2023 年 64 巻 8 号 p. 1724-1738
Low-temperature superplasticity (LTS) is crucial to reduce manufacturing cost and to enhance the applications of superplastic forming. It is well known that grain refinement is the key point to decrease the temperature at which superplasticity is attained. Therefore, ultrafine-grained (UFG) materials have become attractive for achieving LTS. Severe plastic deformation (SPD) techniques provide abnormal grain refinement, and thus they have been used to achieve LTS in metallic materials. This paper overviews and examines the reports of LTS in the severely-deformed metallic materials. It provides fundamentals of grain refinement via different SPD techniques in various classes of metallic materials including Al-, Mg-, and Ti-based alloys. It also gives a brief summary about the effect of microstructural requirements on LTS with an emphasis on grain size, type and chemical composition of grain boundaries and microstructural alteration during the superplastic deformation. In the last section of the manuscript, the main deformation mechanism of LTS were also explained.
Low-temperature superplastic elongations for the severely-deformed alloys.
The ductility of most engineering materials can be increased to desired level that is required for sheet metal forming processes in the machinery and automotive industry after some thermal and/or thermomechanical processes or the addition of alloying elements. On the other hand, as complex shaped parts are needed to be produced especially in aerospace, automotive industries and architectural fields, the ductility of materials should be increased further. Studies have shown that some polycrystalline materials can exhibit extremely high elongation to failure if certain microstructural and experimental requirements are fulfilled.1,2) Superplastic forming methods based on this phenomenon called as superplasticity make it possible to produce very complex-shaped parts in those areas.3)
Superplasticity is defined formally as high neck-free tensile elongation that is achieved in some polycrystalline materials. However, as also stated above, some microstructural and experimental requirements should be fulfilled to attain superplasticity in metallic materials. First of all, the material should be deformed at high temperature regimes above 0.5Tm (Tm is the absolute melting point of the material). Previous studies have shown that grain boundary sliding (GBS) is the main deformation mechanism that occurs during superplastic deformation.4–6) Therefore, grain size of the materials should be decreased down to micron or submicron levels so as to increase grain boundary area, and thus GBS occurs more effectively as the main deformation mechanism leading to high superplastic elongation. Although there isn’t an exact grain size value that must be reached to achieve superplasticity, previous studies have shown that grain size of the metallic materials should be lower than 10 µm.1,7) Besides these two requirements of superplasticity, as the third one, the deformation should be performed at relatively low strain rates ranging between 1 × 10−5 s−1 and 1 × 10−3 s−1 for giving enough time to diffusional deformation mechanism.
It has been well established that two of the aforementioned requirements of superplasticity namely high temperature and low strain rate can be controlled by the third one, i.e., the grain size. High temperature plastic flow, including superplasticity, can be expressed by the constitutive creep equation in the following form:8)
| \begin{equation} \dot{\varepsilon} = \frac{ADGb}{kT}\left(\frac{b}{d} \right)^{p}\left(\frac{\sigma}{G} \right)^{n} \end{equation} | (1) |
Achieving superplasticity at high temperatures and at very low strain rates is considered to be the most important disadvantages of superplastic forming as compared to the conventional forming processes.10) Because, high temperature and low forming rate bring about higher energy requirement to heat up the material to the forming temperature and keep it at that temperature during the manufacturing process. In particular, manufacturing of each part at low strain rates takes approx. 20–30 min which means that the material should be kept at such a high deformation temperature for that long duration.11,12) Thus, regarding that energy consumption constitutes most of the overall cost in manufacturing, any attempt to decrease temperature and increase strain rate at which superplasticity achieved is very crucial to enhance the applications of superplastic forming. Lower forming temperature provides some other benefits which include less damage on forming tools and better surface quality of the formed component. It also prevents severe grain growth, and reduce the cavity formation and solute loss from the surface layer which brings about better post-forming properties.13) Therefore, numerous studies have been performed on the HSRS and LTS, and these two types of superplasticity have been one of the main topics among the material scientists working on superplasticity.
As also stated above, grain refinement is the key point for HSRS and LTS. Thermo-mechanical processes including rolling, forging and extrusion are used to achieve superplastic alloys for industrial superplastic forming applications. On the other hand, the grain sizes of these materials are generally in the range between ∼2–5 µm,1) and thus it is difficult to attain ultra-fine grained (UFG) microstructure (less than 1000 nm) using these conventional plastic deformation techniques. It is well documented that some novel grain refinement techniques based on imposing very high strains to the material via introduction of severe plastic deformation (SPD) provides the capability of decreasing grain sizes down to micron or sub-micron range. Equal channel angular pressing (ECAP),14,15) high pressure torsion (HPT),16,17) friction stir processing (FSP),18,19) accumulative roll bonding (ARB)20) and multi-axial forging (MAF)21) are some of the main SPD techniques. Principles of most commonly used SPD methods and historical development of UFG materials besides with the main properties of nanostructured SPD materials were well-explained in Refs. 21, 22). These methods have been used efficiently to attain significant grain refinement and thus to achieve superplastic behavior in metals at lower temperatures than in conventional superplastic alloys.
The first report on LTS of metals were published in 1988 by Valiev and his co-workers.23) Their work has a special significance since the term LTS was introduced for the first time as “Low Temperature Superplasticity of Metallic Materials”. Valiev et al.23) achieved an elongation of 250% in a severely-deformed UFG Al–Cu–Zr with a grain size of 300 nm. Achieving such a high elongation at only 220°C corresponding to 0.53 Tm which is relatively lower than the temperature of conventional superplasticity led to the studies on a new type of superplasticity so-called LTS. Particularly achieving superplasticity in the nanocrystalline nickel at the lowest normalized superplastic temperature of 0.36Tm24) provided a driving force to the LTS-related studies. Ideal LTS refers to the superplastic deformation that is achieved at low homologous temperatures below 0.5 Tm.25,26) On the other hand, particularly LTS studies on the Al-based alloys have been focused on the temperatures ranging between 423 K to 623 K corresponding to 0.45–0.64 Tm of the Al alloys.23–43) Similarly, superplastic elongations achieved in the Mg-based alloys below 573 K (∼0.62 Tm) were considered as LTS in the open literature.44–60)
In view of above, the present study overviews and examines the reports of LTS in severely-deformed metallic materials. The following section describes the UFG formation in metallic materials by SPD to attain LTS and summarizes the achieved superplastic elongations. In the next part, room temperature superplasticity (RTS) was summarized. In the last two parts deformation mechanisms operated at LTS and innovation potential of LTS were explained.
Table 1 summarizes the studies aiming to produce UFG microstructure in aluminum alloys by various SPD techniques to achieve UFG microstructure leading to LTS. In order to make more precise comparison, it is reasonable to categorize the samples according to chemical composition rather than the utilized SPD processes, and to compare the achieved results. It is clear that 7xxx series Al alloys containing Zn and Mg as the major alloying elements are one of the most commonly studied ones for LTS.25–43) As the SPD processes, HPT,27,30–32) FSP25,26,28) and ECAP29) have been successfully used for grain refinement to attain UFG microstructure in this alloy system. In general, HPT brought about lower grain sizes in 7xxx series Al alloys as compared to two other SPD techniques. The lowest grain size was reported to be 120 nm after HPT process applied to Al7075 alloy at RT and under a pressure of 6 GPa for 3 turns with a rotation speed of 1 rpm.27) Other HPT processes also resulted in extremely low grain sizes below 200 nm in 7xxx series Al alloys.30–32) Particularly, detailed microstructural examinations of Chinh et al.32) on the HPT-processed 7075 alloy showed that Zn and Mg atoms segregated at the matrix grain boundaries. As will be described later in this study, segregation of Zn atoms at the grain boundaries is favorable for LTS since it contributes the GBS as the main deformation mechanism.32) Regarding the studies where the FSP was used as the grain refinement tool, the effect of cooling of the anvil on the final grain size of FSP-processed samples were investigated. Orozco-Caballero et al.28) performed FSP process on the Al 7075 alloy using two different backing anvils. First one was a conventional martensitic stainless steel without cooling, and the second was a copper anvil cooled with liquid nitrogen. It was revealed that grain size of the sample processed on the cooled anvil was 300 nm which is relatively smaller than that of the sample processed on the uncooled one. Similar result was also reported in an another study by Liu and Ma.25) They used room-temperature water to quench the sample immediately behind the FSP tool to prevent the growth of the recrystallized grains during the FSP process. It was stated that the achieved grain size of 800 nm was significantly smaller than that of obtained by a conventional FSP process without cooling. ECAP was not considered much as a grain refinement tool in 7xxx series Al alloys to achieve LTS. The only study where ECAP was utilized to attain LTS in 7xxx series Al alloys was performed by Malek et al.29) In that study, Al 7075 alloy was subjected to 6 passes ECAP at 120°C following route Bc, and grain size of the alloy was reported to be 500 nm after that process.
Al–Mg alloys were also considered excessively for LTS into Al-based alloys.33–40) As the SPD processes, FSP, ECAP, ARB, HPT and MAF processes were used to achieve LTS in this alloy system. Ma and Mishra33) investigated the effect of FSP tool size on the final microstructure of the Al–4Mg–1Zr alloy. They used two different tool sizes in which the pin and shoulder diameter of the second tool were reduced by 50% and 33%, respectively, as compared to the first one. Furthermore, tool transverse-speed/rotation rate of the first tool was kept higher than that of the second one. Although the sample processed with the first tool having higher tool transverse-speed/rotation rate was expected to have lower grain size due to low thermal input,61–63) it showed higher grain size than the sample processed with the tool having smaller sizes. It was concluded that tool size was the dominant parameter that determines the size of recrystallized grains. Decreasing tool size reduced the thermal input due to the less contact area between the tool and work piece and resulted a UFG microstructure with 700 nm grain size. In an another study, Liu and Ma34) processed an Al–Mg–Sc alloy via FSP with immediate water quenching of the samples just behind the FSP tool. Similar to the results reported for Al7075 alloy,25) inhibiting the growth of the recrystallized grains by cooling via water quenching immediately behind the FSP pin tool resulted a UFG microstructure with an average grain size of 600–700 nm. ECAP was also utilized to achieve UFG microstructure resulting in LTS in Al–Mg alloys. It was applied to three different Al–Mg alloys in various studies following the same ECAP route (route Bc) and using die with the same channel intersection angel of 90°. Regarding these studies, the effect of processing temperature on the final grain size of the samples can clearly be seen.36,37) The sample processed up to 8 passes at RT exhibited the finest grain size of 200 nm36) while grain size of the sample subjected to 16 passes ECAP at 598 K was reported to be 1000 nm.37) Actually these results are well consistent with well-known evidence of the dependence of final grain size on the ECAP temperature that the finest grain size would be attained when the pressing is performed at the lowest possible temperature.64) MAF was an another SPD method used to attain LTS in Al–Mg alloys. Noda et al.39) processed Al5083 alloy with MAF at 543 K up to 10 cycles in order to impose a true strain of ∼6 to the sample. The mean grain size was reported to be less than 1000 nm after that process. The ARB was also used successfully to get UFG microstructure in an Al–Mg alloy. Five cycles of ARB which corresponds to a total true strain of 4 resulted in a UFG microstructure in Al5083 alloy with a mean grain size of 280 nm.40)
Besides 7xxx series and Al–Mg alloys, Al–Mg–Si (6xxx series), Al–Mg–Li and Al–Cu alloy systems were rarely studied to achieve LTS. Seven cycles of ARB brought about 200 nm grain size in Al6063 alloy.41) Mishra et al.42) processed Al–Mg–Li alloy (Al1420) via HPT to impose a true strain of 6 at RT and under a high pressure of 1.2 GPa. This process led to the smallest grain size in all severely-deformed Al alloys exhibiting LTS, and the mean grain size of the Al–Mg–Li alloy was reported to be 100 nm.42) Imposing a large amount of strain (∼15–20) to Al–4Cu–0.5Zr via ECAP performed at 473 K also resulted in an extremely low grain size of 150 nm.43)
Regarding the results summarized above the following conclusions can be made about the SPD processing of Al alloys to attain UFG microstructure required for the LTS. It is clear that HPT is the most effective way to achieve UFG microstructure having grain size lower than 200 nm even though it is applied for low revolution number.27,30–32,42) It is not much reasonable to compare the achieved grain sizes directly since the HPT and other SPD methods were applied to the Al alloys at different temperatures and for various imposed-strain levels. However, previous studies also showed that HPT provides more effective grain refinement than ECAP just in the case of Al alloys included in the present study.65–69) HPT process is performed under high pressure which has a significant effect on the microstructural homogeneity across the process sample and also grain refinement.16,70) It was shown that higher pressure brings about more homogeneous microstructure and also higher hardness value besides with the smaller grain size.16,70) It reasonable to concluded that besides with the imposed strain, applied pressure is an extremely important parameter dictating the extent of microstructural refinement during HPT processing as compared to the ECAP.16,70) ECAP, on the other hand, also provides opportunity to attain UFG microstructure in Al alloys.29,36–38,43) However, it should be conducted at the lowest possible temperature where the samples could be processed without any crack formation.64) In general, grain sizes of FSP-processed samples25,26,33–35) are higher than the samples processed via HPT27,30–32,42) and ECAP.29,36–38,43) On the other hand, it was seen that cooling the FSP anvil in any manner or water quenching of the sample just behind the FSP tool result in reasonable small grain size. Similar results were also reported in some other FSP-processed Al alloys.71,72) It was shown that the typical recrystallized grain sizes of the FSP-processed aluminum alloys processed without cooling is in the micron range while UFG microstructures having grain size lower than 1 µm have been achieved by using external cooling.71,72) Using the cooled anvil results in lower peak temperature during the FSP process and brings about lower grain size due to the well-known principles of recrystallization where smaller grains are achieved at the lower annealing temperature.73) Cooling the sample just behind the FSP tool, on the other hand, could restrain growth of the recrystallized grains during the FSP process and contribute to attain smaller grain sizes. ARB and MAF are two other SPD methods used to obtain UFG microstructure in Al alloys, and considerably lower grain sizes were reported after these processes.
2.1.2 LTS in severely-deformed UFG Al-based alloysLow-temperature superplastic elongation values of the severely-deformed Al-based alloys are shown in Fig. 1. It should be taken into consideration that the elongation to failure values were obtained from the tensile test samples having various dimensions in different studies. It is well known that sample dimensions have a direct effect on the ductility and measured elongation to failure tends to increase when the cross-sectional area of the tensile test samples are increased.74) Therefore, it should be avoided to compare the superplastic elongations given in Fig. 1 to 5 regardless of tensile test samples’ dimensions. The results on the 7xxx series and other Al-based alloys including Al–Mg system are given separately in Fig. 1(a) and 1(b), respectively, so that the graphs in the figures can be seen clearly. Among 7xxx series alloys, the highest elongation was achieved as 690% in the HPT-processed Al7050 at 473 K (corresponding to the 0.52 Tm) due to the UFG microstructure of the alloy.30) Regarding all severely-deformed Al-based alloys, ECAP-processed Al–4Cu–0.5Zr reflected the highest low-temperature superplastic elongation as 850% at 523 K (0.56 Tm).43) Besides these two studies, all of the superplastic elongation values of other Al alloys given in Table 1 and Fig. 1 change between 220%40) and 590%.35)
Superplastic elongation values achieved in severely-deformed; (a) 7xxx series and (b) other Al-based alloys besides with the grain sizes and test temperatures.
The lowest homologue temperature at which the low-temperature superplasticity was achieved in Al-based alloys was reported to be 0.45 Tm.31,32) It was reported that Al–4.8Zn–1.2Mg–0.14Zr alloy processed via 10 turn HPT exhibited a superplastic elongation of 400% at a quite low temperature of 423 K.31,32) Besides the UFG microstructure with a mean grain size of 200 nm, Zn segregation at the Al/Al grain boundaries were stated to be the main reason of lowering the temperature at which the superplasticity was attained. Zn and partial Mg segregation at the grain boundaries made the Al–Zn–Mg–Zr alloy to be considered as the two-phase system containing a grain interior phase and a thin grain boundary layer enriched in Zn and Mg. Since Zn and Mg addition decreases the melting point of Al,32) melting point of the grain boundary region could be decreased more than 100 K as compared to the grain interior. Thus, even the alloy deformed at 423 K, it corresponded to an elevated homologue temperature higher than 0.5 Tm for the grain boundaries. It is worth to point out that Mg segregation at the grain boundaries has a negative effect on GBS.75) It was shown that UFG Al–Mg did not exhibit superplastic elongation without Zn addition.75) On the other hand, Zn segregation at the grain boundaries in Ref. 32) resulted in a grain boundary activation energy of ∼70 kJ/mole which is significantly close to the activation energy of superplastic deformation for the Al–4.8Zn–1.2Mg–0.14Zr alloy. Thus the segregation of Zn at the grain boundaries accelerated diffusion and enhanced sliding which brought about superplasticity at lower temperatures than the other Al-based alloys.31,32)
From the results summarized above it is clear that grain boundary character in the microstructure has also a significant effect on the temperature of superplasticity. GBS occurs easily at high angle grain boundaries (HAGBs) while low angle grain boundaries (LAGBs) are considered as immobile regarding the GBS.76) It was reported that contribution of GBS to total elongation exceeded 50% even at 443 K in the FSP-processed Al–Mg–Sc alloy with the fraction of HAGBs as high as 95% due to the easier GBS.34) It was concluded that UFG microstructure with predominantly HAGBs that could resist intergranular separation, and significant grain growth during the superplastic deformation would be beneficial to achieve LTS.35) Besides the grain boundary character, dynamic recovery leading to balance of work hardening rate and thus delay necking was stated to be an another origin of LTS in the UFG Al6063 processed via ARB.41) It was also pointed out that elongated grains of the ARB-processed alloy transformed into equiaxed ones, and the fraction of the HAGBs significantly increased during deformation due to the dynamic recovery/recrystallization. Such microstructural alteration had also a contribution to LTS.
2.2 Mg-based alloys 2.2.1 UFG formation in Mg-based alloys for LTSLTS related studies on the severely-deformed Mg-based alloys are summarized in Table 2 besides the utilized SPD techniques and achieved results including grain size, superplastic elongation, m-value, deformation temperature etc. AZ and ZK series Mg alloys are two of most commonly considered ones to attain LTS.44–54) These alloys were subjected to ECAP44–46,49–54) and HPT47,48) processes to achieve fine grained (FG) and UFG microstructures required for the LTS. Mabuchi et al.44) applied ECAP to an as-extruded AZ91 alloy at 448 K to impose a total strain of ∼6.4. Some of ECAP-processed samples were subjected to annealing at 498 K for 12 h to convert non-equilibrium grain boundaries to equilibrium ones to examine the effect of grain boundary type on the superplastic behavior of the alloy. Grain sizes after ECAP and ECAP + annealing processes were reported to be 700 nm and 3.1 µm, respectively. A similar process was also applied to the extruded AZ91 in an another study.45) Post-ECAP annealing of the alloy that was exposed to ECAP for ∼8 strain resulted in an average grain size of about 1 µm.45) Six passes ECAP of the as-cast AZ91 alloy at 473–523 K following route Bc brought about almost the same grain size with the Ref. 45), and mean grain size was achieved as 800 nm after that process.46) HPT process was also applied to AZ91 alloy at RT up to 10 turns in two different studies.47,48) Grain sizes of the alloy after these processes were not specified. However, it can be concluded that HPT process resulted in a UFG microstructure with a grain size less than 1 µm regarding the grain sizes of the tensioned samples provided in Ref. 48). As an another AZ series alloy, initially extruded AZ31 was subjected to 8 passes ECAP at 473 K via route Bc leading to an average grain size of 700 nm.49) These results indicate that ECAP processing of the AZ series Mg alloys brings about UFG microstructures having grain sizes in the range of 700–1000 nm that would be suitable for LTS.
ECAP was used successfully to attain UFG microstructure and LTS in ZK60 and ZK40 alloys. Yu et al.50) applied ECAP processes to an as-extruded ZK60 alloy at three different temperatures of 453 K, 473 K and 493 K using an ECAP die that was capable of applying a back pressure of 40 MPa during the ECAP process. It was reported that 4 passes ECAP at 473 K resulted in a smaller grain size (1.2 µm) than that of performed at the lower temperature of 453 K (1.6 µm).50) Even though this result is not consistent with well-known relationship between final achieved grain size and process temperature of ECAP, detailed analysis was not provided in that study to explain that paradox. In an another study, as-extruded ZK60 alloy was subjected to 8 passes ECAP at 433 K followed by a post-ECAP annealing at 473 K for 30 min.51) This two-step thermomechanical process led to formation of a microstructure with a mean grain size of 1.4 µm. Mostaed et al.52) processed as-extruded ZK60 alloy via a two-step ECAP where 4 passes were performed at 523 K followed by 4 more passes at 423 K. The mean grain size was reported to be 500 nm after this process. ECAP was also applied to the ZK60 alloy in the as-cast condition.46,53) Six passes ECAP that was applied to as-cast ZK60 alloy at 453–503 K resulted in an average grain size of 1.0 µm in Ref. 46). Lapovok et al.53) also processed as-cast alloy successfully via ECAP performed at 473 K up to 6 passes using route Bc. It was stated that this process led to formation of a bi-modal microstructure having grain sizes in the range of 1.8 ± 1.5 µm for small grains and about 12.5 ± 2.2 µm for coarse grains.53) ZK40 was also processed with the ECAP process, and sub-micron grain-sized microstructure was reported in the as-pressed alloy.54)
Mg–Li55–57) and Mg–Al58,59) alloys were also studied to attain LTS. Zhou et al.55) processed Zn–9Li–1Zn alloy via FSP using quite low tool rotation speed of 30 rpm and a processing speed of 10 mm/min. It was stated that extremely low heat input due to low rotation speed of 30 rpm resulted in a UFG microstructure with an average grain size of 610–960 nm. Furthermore, FPS distributed α- and β-phases uniformly and increased the length of α/β boundaries that would be beneficial to enhance the superplasticity of the alloy due to the high diffusivity along these boundaries.55) A two-step ECAP process involving an annealing stage between two steps of ECAP was applied to Mg–10Li–1Zn alloy by Yoshida et al.56) The first step of ECAP performed at 323 K and subsequent annealing at 623 K for 1 h were designed to refine and spheroidizate the α phase of the as-cast alloy. Four more passes of ECAP were applied to the alloy at 323 K to attain further grain refinement. The final grain size was not specified in this study, but the achieved microstructure was stated to be favorable to attain LTS in the Mg–10Li–1Zn alloy.56) As an another Mg–Li alloy, as-extruded Mg–8Li was processed via HPT that was utilized at RT and under a pressure of 3 GPa up to 5 turns.57) It was reported that HPT resulted in a smaller grain size of ∼500 nm than pure Mg due to the effect of solute atoms on the mobility of dislocations.57) Matsubara et al.58) processed as-extruded Mg–9Al alloy with 2 passes ECAP performed at 673 K using route Bc. The mean grain size of the ECAP-processed alloy was reported as 700 nm. Effect of pre-HPT extrusion step on the microstructure and mechanical properties of Mg–9Al alloy was investigated by Kai et al.59) They applied 5 turns HPT process performed at 423 K to two different alloy samples namely as-cast and as-extruded. As-cast + HPT process brought about a slightly finer grain size of 330 nm as compared to extrusion + HPT process yielding an average grain size of 370 nm. They also investigated the effect of HPT temperature on the microstructure and superplasticity of Mg–9Al alloy. It was showed that the sample processed via HPT performed at RT exhibited lower mean grain size of 150 nm than that of the sample processed at 473 K. However, some limited cracks were observed on the sample processed at RT which restrained to achieve as high as expected superplastic elongation in that sample.59)
As an another Mg-based alloy, ZM21 was processed via a two-step ECAP process including 4 passes ECAP at 473 K followed by 4 more passes at 423 K.60) It was reported that, the first step ECAP process brought about remarkably refined microstructure with an average grain size of 700 nm. Second step ECAP, on the other hand, resulted a slight grain growth and grain size reached to 900 nm due to the relatively longer exposition to high processing temperature.60)
Considering the studies summarized above some important conclusions can be made about the ECAP processing of Mg-based alloys to attain FG/UFG microstructures required for LTS of these alloys. First of all, in most cases, ECAP process was preferred to be applied to the Mg alloys following a pre-ECAP process to avoid any crack formation that may occur during the ECAP due to the limited slip system in the HCP structure. Mostly extrusion was considered to eliminate the as-cast microstructure and enhance the ductility prior to ECAP process.44,45,50,51) Second, an ECAP facility that is capable of applying back pressure was also used to prevent the samples from crack formation.50) It has been well-established that back-pressure has an important role on the workability of metals during ECAP. It was shown that 16 passes of ECAP could be applied to a brittle Al–5% Fe alloy in the presence of back-pressure while the alloy broke after only one or two passes in the absence of back-pressure.77) Furthermore, using a back pressure decreased the temperature at which AZ31 alloy processed successfully via ECAP from 473 K to 373 K.78) It was stated that nucleation and growth of cracks was suppressed by a compressive hydrostatic pressure due to the back pressure.78) Thus, it can be concluded that using back pressure should be considered as an important ECAP parameter to process difficult-to work Mg-based alloys successfully to attain UFG microstructure and LTS. Finally, high pressing temperature was selected to process the Mg-alloys successfully without cracking. Almost all ECAP processes were performed at the elevated temperatures and most frequently the temperature was above 448 K.44–46,49–54) It is clear that ECAP temperature has a direct effect on the final grain size of the Mg alloys, but it is not possible to make a generalization between the achieved grain size and ECAP temperature. Even though lower temperature is expected to result in finer grain size,64) relatively smaller grain size was reported in ZK60 alloy processed at 473 K as compared to a counterpart that was subjected ECAP with the same process parameters but at a lower temperature of 453 K.50) Contrary to the ECAP, HPT process was applied to Mg-based alloys successfully even at RT because of high processing pressure.57) HPT resulted in a more refined microstructure in the Mg alloys as compared to the other SPD techniques due to the high imposed strain and low processing temperature besides with the high processing pressure.47,57,59)
2.2.2 LTS in severely-deformed FG/UFG Mg-based alloysFigure 2 shows the superplastic elongations attained in the Mg-based alloys at low temperatures. Looking at Fig. 2, Mg–8Li alloy having a mean grain size of 500 nm revealed 400% elongation at 0.37 Tm (323 K) that is the lowest homologue temperature at which LTS was achieved in the Mg-based alloys.57) About 700% elongation was also reported in that sample tested at 373 K and at a strain rate of 1 × 10−3 s−1.57) The highest elongation, on the other hand, was reported as 2040% in the HPT-processed ZK60 alloy tested at 493 K corresponding to 0.54 Tm of the alloy. Such a high elongation was attributed to the bi-modal microstructure of the alloy.53) HPT-processed AZ91 alloy also showed a considerably high elongation of 1190% at 473 K.47) Furthermore, this sample exhibited superplastic elongation at a relatively low temperature of 423 K which is equal to 0.47 Tm of the AZ91.47) Thermally stable UFG microstructure with equiaxed grainy morphology was stated to be the main reason of the achieved superplasticity. Furthermore, nano-sized β-phase (Mg17Al12) particles that was dispersed mainly along the grain boundaries also contributed to the superplastic elongation by inhibiting grain growth and accommodating GBS as the main deformation mechanism.47) Contribution effect of precipitates on the GBS was also reported for Mg–10Li–1Zn and Mg–9Li–1Zn alloys.55,56) It was stated that β-phase (lithium solid solution) particles precipitated at the grain boundaries and triple junctions contributed the stress relaxation and accommodated GBS.56) It was also specified that Li segregation at α/α phase boundaries (α is the Mg solid solution) increased the boundary diffusivity and contributed phase/boundary sliding at relatively low temperature leading to high superplastic elongation of 1104% in Mg–9Li–1Zn alloy.55) Another reason of such a high elongation of 1104% was stated to be increment in the length of α/β boundaries which have higher diffusivity than the α/α boundaries. Thus increasing the length of α/β phase boundaries contributed to the GBS and led to high superplastic elongation.55) These results show that diffusivity along the grain boundaries plays a significantly important role on the LTS in Mg–Li based alloys. Equation (1) also shows such an effect of grain-boundary diffusion on the temperature at which superplastic elongation is attained. According to the eq. (1) low-temperature superplasticity can be achieved by enhancing grain-boundary diffusion. On the other hand, it is well known that diffusion coefficient depends strongly on the deformation temperature and higher temperature brings about higher diffusion coefficient. Thus, diffusion coefficient should be increased without increasing temperature in order to achieve LTS. Results of the Mg–Li–Zn55,56) and Al–4.8Zn–1.2Mg–0.14Zr31,32) alloys show that diffusivity along the grain boundaries can be increased by modification of the chemical composition of grain boundaries. Li segregation at the Mg–Li–Zn and Zn segregation at the Al–4.8Zn–1.2Mg–0.14Zr alloys increase the boundary diffusivity and contribute phase/boundary sliding leading to LTS. Effect of grain-boundary segregation on the mobility of grain boundaries led to achieve even RTS in Al–30Zn79) and Mg–8Li80) alloys. The details of these studies are given in sections 3. However, segregation of Li and Zn atoms at the grain boundaries requires large number of HPT cycles.79,80) All these results show that besides with the UFG formation, controlling the grain boundary composition due to the grain-boundary segregation by means of proper SPD method and using appropriate alloying elements (such Li, Na, Ca, Sr, Se, Bi and Te to Mg which can enhance the diffusivity)79,81) would be considered as a key point to achieve LTS.
Superplastic elongation values of the severely-deformed Mg-based alloys besides with the grain sizes and test temperatures.
It is reasonable to mention that the ECAP-processed + annealed AZ91 alloy with a mean grain size of 3.1 µm showed 956% elongation which is quite higher than that of achieved in the ECAP-processed alloy having considerable lower grain size of 700 nm.44) Regarding that the samples with the same dimensions were subjected to tensile tests at the same temperature of 473 K, the sample with lower grain size would be expected to reflect higher superplastic elongation.44) It was stated that such an important difference between the achieved superplastic elongations in ZK60 alloy arose from the grain boundary characteristics of the samples. Non-equilibrium grain boundaries of the ECAP-processed sample were converted to equilibrium ones by annealing the sample at 498 K for 12 h. It was shown that non-equilibrium gran boundaries had long-range stresses and many regular or irregular arrangements of facets and steps.44) Thus, it was concluded that dislocation slip for accommodation of GBS were inhibited by the long-range stresses at non-equilibrium grain boundaries which brought about lower superplastic elongation in the ECAP-processed sample.44) Watanabe et al.51) also confirmed the importance of equilibrium grain boundaries to attain LTS in Mg-based alloy. It was shown that 1083% superplastic elongation was achieved in the ECAP-processed + annealed ZK60 alloy with a mean grain size of 1.4 µm at 473 K.51) In an another study, 500 nm grain-sized ECAP-processed ZK60 alloy, on the other hand, showed a superplastic elongation of 420% at the same temperature.52) Knowing that grain boundaries in fine-grained materials subjected to SPD are often in non-equilibrium state,44,53) the differences between the achieved superplastic elongations in these two studies of Refs. 37, 51, 52) could be explained by the type of the grain boundaries. These studies indicate that equilibrium grain boundaries are favorable to attain high low-temperature superplastic elongations in Mg-based alloys.
2.3 Ti-based alloys 2.3.1 UFG formation in Ti-based alloys for LTSRelated studies on LTS of Ti-based alloys processed via SPD methods are summarized in Table 3. It is seen that almost all of the studies were performed on the Ti–6Al–4V alloy for this purpose82–90) while some limited studies were conducted on other Ti-based alloys.91–93) Ti–6Al–4V was processed via FSP by the same researchers in two different studies using a processing speed of 30 mm/min and tool rotation speed of 120 rpm.82,83) The mean grain sizes of the FSP-processed alloy were reported as 570 nm and 510 nm in these studies. HPT was utilized to attain UFG microstructure in Ti–6Al–4V alloy by Zhang et al.84) Prior to HPT, they applied a heat treatment at 1283 K for 1 h and at 823 K for 3 h followed air cooling in order to achieve a fully lamellar structure and investigate the effect of such an initial microstructure on the final grain size of alloy. It was stated that 20 turns of HPT performed under a pressure of 6 GPa resulted in a significant grain refinement and mean grain size was measured to be 70 nm. Increasing HPT turn number to 30 brought about further grain refinement and the average grain size was reported as 50 nm.84) It was concluded that final grain size of the HPT-processed Ti–6Al–4V alloy depends strongly on the volume fraction of the lamellar structure in the pre-HPT microstructure. It was also stated that higher volume fraction of the lamellar structure resulted in more significant grain refinement during HPT.94) The same pre-HPT annealing and HPT processing parameters were used to process Ti–6Al–4V alloy, and similar results were also reported in an another study.85) Shahmir et al.86) also investigated the effect of initial microstructure on the final grain size and superplasticity of Ti–6Al–4V. They used two different samples, the first one was solution treated and quenched while second one was solution treated and cooled in air. It was shown that quenched sample exhibited finer grain size of 30 nm than the air-cooled one (40 nm) after HPT processing. It was stated that martensitic microstructure of the quenched sample contains high level of residual stress, many dislocations, stacking faults and high volume fraction of grain boundaries which probably act as nucleation sites for rapid grain fragmentation and subgrain formation and promote grain refinement.86)
As an another SPD method, Ko et al.87,88) applied 4 and 8 passes ECAP to the Ti–6Al–4V alloy at 873 K. Regardless the number of ECAP, the grain size was achieved to be 300 nm. However, increasing the ECAP pass number resulted in an increment in the misorientation angles of grain boundaries.88) MAF was also used to achieve UFG microstructure in Ti–6Al–4V alloy. Kral et al.89) processed the alloy via warm MAF at the temperature interval of 973 K and 748 K, and this process yielded a mean grain size of 150 nm. MAF with the same parameters in Ref. 89) was also applied to Ti–6Al–4V alloy, and it was followed by a rolling step performed at 748 K for a total reduction of 50%. The mean grain size of the alloy was reported to be 100–400 nm after this two-step thermomechanical process.90)
Besides the Ti–6Al–4V, some other Ti-based alloys were also considered for LTS.91–93) Ti-15-3 (Ti–5Al–5V–5Mo–1Cr–1Fe) alloy was processed by the same research group via FSP in two different studies.91,92) Zhang et al.91) performed FSP process with a tool rotation speed of 250 rpm and tool transverse speed of 100 mm/min. They also applied argon shielding gas to prevent the processed surface from oxidation. It was shown that FSP process decreased the grain size of the alloy from 44 µm to 6.6 µm.91) In an another study, they decreased tool rotation speed to 100 rpm while tool transverse speed was kept constant as 100 mm/min.92) Relatively fine β-phase with a mean grain size of 1.6 µm was achieved in that study due to the ultra-low heat input as a result of slow rotation speed of 100 rpm. It was concluded that the peak temperature in the stir zone was lower than the β-transition temperature which makes easier the precipitation of a small amount of α phase.92) Thus, the pinning boundary effect of the α-phase and low temperature input resulted in relatively smaller grain size in the Ti-15-3 alloy. Ratochka et al.93) applied MAF to the near β-titanium alloy of Ti–5Al–5V–5Mo–1Cr–1Fe to produce UFG microstructure and attain LTS. MAF process brought about a significant grain refinement, and grain/subgrain size was reported to be 170 nm which is quite close to that of achieved in Ti–6Al–4V that was processed via the same SPD method.89)
Considering the results of the studies summarized above, the lowest grain size was achieved after HPT processing of Ti-alloys.86) Furthermore, the initial microstructure has a significant effect on the grain refinement that occurs during the HPT processing of Ti–6Al–4V alloy. Particularly, formation of martensitic microstructure or increasing the volume fraction of lamellar structure before the HPT would be beneficial to attain finer grain size.84,86) Grain sizes of Ti-based alloys processed via FSP are slightly larger than that of processed using other SPD methods.82,83,91,92) However, decreasing tool rotation speed would be favorable to achieve FG microstructure since it results in relatively low heat input during the FSP process.92)
2.3.2 LTS in severely-deformed FG/UFG Ti-based alloysSuperplastic elongation values achieved in Ti–6Al–4V as the most commonly studied Ti-based alloy for LTS are summarized in Fig. 3 besides with the grain sizes and test temperatures. It would be more reasonable to compare the results achieved at each test temperature in itself. First of all, the highest superplastic elongation was achieved as 1400% in the FSP-processed alloy having a grain size of 570 nm.82) Tensile tests were performed at an initial strain rate of 3 × 10−3 s−1 and a temperature of 873 K corresponding to 0.45 Tm of the alloy (Fig. 3 and Table 1). Such a high superplastic elongation at considerably low temperature of 873 K was attributed to the microstructural characteristics of the FSP-processed sample. First of all, it was stated that uquiaxed UFG microstructure with high amount of HAGBs (90.7%) contributed to the superplastic elongation since HAGBs have higher grain boundary energy as compared to LAGBs leading to much easier GBS. It was also stated that presence of β-phase precipitates at α-phase boundaries provided two benefits to enhance the superplastic elongation. First, volume fraction of β-phase increased during the deformation which is beneficial for superplasticity since GBS along the heterogeneous α/β-phase boundaries is much easier than those of homogeneous boundaries.82) Second, β-phase precipitates inhibited the excessive grain growth during the tensile test and promoted the achievement of the excellent LTS.82) In an another study, HPT-processed samples having extremely low grain sizes of 70 nm and 80 nm and thermally stable microstructure with high fraction of HAGBs showed an elongation to failure of less than 800% at 873 K.85) Regarding their extremely low grain sizes achieved superplastic elongation values are quite lower than that of reported in Ref. 82) at the same temperature. These results seem to be not consistent with well-known relationship between the grain size and superplastic elongation that decreasing grain size of a superplastic material brings about a higher elongation to failure.1–3,7) Since their microstructural features are almost same, such a paradox between grain size and elongation to failure could be explained regarding dimensions of the tensile test samples used in these two different studies. As stated above larger cross-sectional areas of the tensile test samples bring about higher measured elongation to failure.74) While the cross-sectional area of the FSP-processed sample was 1.7 mm2,82) it was only 0.5 mm2 for the HPT-processed one.85) Thus, while the HPT-processed sample had extremely low grain sizes of 70 nm and 80 nm, it exhibited relatively low superplasticity as compared to the FSP-processed sample due to the reasonably smaller cross-sectional area of its tensile test sample. Another interesting result in Fig. 3 is that two ECAP-processed samples having the same grain sizes of 300 nm showed relatively different superplastic elongations at the same temperature of 973 K.87,88) The sample processed via 8 passes ECAP exhibited an elongation to failure of 700%88) while the elongation value was 474% for the sample processed with 4 passes of ECAP.87) It was stated that increasing ECAP pass number from 4 to 8 resulted in an increment of the misorientation angle of grain boundaries and also decreased the activation energy.88) Thus, increasing the imposed strain level was found to be beneficial to improve superplastic elongation due to the enhanced activation of GBS. HPT-processed sample having an extremely low grain size of 30 nm showed a superplastic elongation of 815% at 1 × 10−3 s−1 86) which was reasonably higher than those of the ECAP-processed samples tested at the same temperature of 973 K.87,88) Taking into consideration the quite small tensile test sample of the HPT-processed sample, it showed a considerable superplastic elongation of 600% at even a high strain rate of 1 × 10−2 s−1.86) MAF + rolled Ti–6Al–4V alloy with a mean grain size of 100–400 nm showed a high superplastic elongation of 1000% at 823 K corresponding to the lowest homologue temperature of 0.43 Tm where the LTS was attained in Ti-based alloys.90) Besides the UFG microstructure, transformation of the morphology of β-phase during the superplastic deformation was stated to be main reasons of high superplastic elongation. It was shown that β-phase initially located at the triple junction and α/α-grain/particle boundaries transformed into a continuous network separating the α-phase particles after the superplastic elongation and provided a wetting effect. Thus soft lubricating layer between α-phase particles contributed to high superplastic elongation.90)
Superplastic elongation values achieved in severely-deformed Ti–6Al–4V alloy besides with the grain sizes and test temperatures.
Severely-deformed Ti–15V–3Cr–3Sn–3Al and Ti–5Al–5V–5Mo–1Cr–1Fe alloys also exhibited high low-temperature superplastic elongations. FSP-processed Ti–15V–3Cr–3Sn–3Al alloy with a mean grain size of 1.4 µm showed a superplastic elongation of 842% at 3 × 10−4 s−1.92) The α-phase precipitation and the continuous dynamic recrystallization (CDRX) during superplastic deformation as well as the fine and equiaxed microstructure with the relatively high ratio of HAGBs and random crystallographic orientations were stated to be the main reasons of the high superplastic elongation.92) MAF-processed Ti–5Al–5V–5Mo–1Cr–1Fe alloy also showed a high superplastic elongation of 950% at a relatively higher strain rate of 2 × 10−3 s−1.93)
2.4 A comparison of LTS in Al, Mg and Ti-based alloysThe highest superplastic elongations as a function of strain rate for Al-, Mg-, and Ti-based alloys using the experimental data in the Table 1, Table 2 and Table 3 are provided in Fig. 4. In general elongation to failure values of Mg-based alloys are scattered in a wide range of 330% and 1330% with an exception of 2400% which is also the highest low-temperature superplastic elongation in severely-deformed alloys. Ti-based alloys, on the other hand, exhibit superplastic elongations between 474% and 1400%. Relatively lower superplastic elongations below 700% were reported in the Al-based alloy. Furthermore, elongations to failure values reported in some studies are less than 400% for Al alloys.25,28,29,33,38,40,41) The ductility of 200% and the m-value of 0.3 are considered to be the bottom limits of superplastic behavior in some studies.19,26,35,95) More recently, on the other hand, Langdon2) suggested that 400% and strain rate sensitivities close to 0.5 should be considered as the evidence for superplastic flow in metals. As can be seen, there is no an exact elongation value that must be achieved to consider a material as superplastic. Therefore, elongation to failure values exceeding 200% were listed as superplastic in the present study.
As stated above, grain refinement is the key point to achieve LTS and almost all of the studies aiming to achieve LTS are based on this principle. It was shown that significant grain refinement brought about superplastic behavior even at RT in some low-melting point alloys. Therefore, it’s thought to be worthy to give the RTS at a separate section.
3.1 UFG formation in the metallic materials for RTSA previous overview paper described the available studies aiming to investigate the RTS in a range of metals and alloys after SPD processing.96) The SPD processes that were used to achieve UFG microstructure, superplastic elongation values and deformation mechanisms were examined in detail in that overview paper. Therefore, the details of the utilized SPD processes to attain RTS were not given in the present study. A brief summary about the fundamentals of SPD processing of metallic materials exhibiting RTS was provided instead. Furthermore, recently published articles were included more comprehensively. Table 4 summarizes the reported data on RTS in the severely-deformed metals.79,80,97–109) Besides being a modal superplastic alloy, eutectoid Zn–22Al is the most commonly studied one to attain RTS. It is seen that almost all studies used ECAP as the grain refinement tool to achieve FG/UFG microstructure in this alloy.97–104) These studies investigated the effects of ECAP temperature, phase regime where ECAP is performed and number of ECAP passes on the achieved microstructure of Zn–22Al. It was shown that ECAP temperature should be kept as low as possible when the alloy is processed in two-phase region (below the etectoid temperature of 275°C).100,101) It was also demonstrated that application of ECAP during the phase transformation of the quenched alloy and in the single phase region are more efficient in grain refinement.97,98) A two-step ECAP-processing where the first step was applied to the alloy in the single phase region followed by a second step performed at RT, on the other hand, brought about the lowest grain size of 200 nm in Zn–22Al alloy.102) Eutectic Zn–5Al and quasi-single phase Zn–0.3Al alloys are the other severely-deformed Zn–Al alloys exhibiting RTS.106,107) As-quenched Zn–5Al alloy was processed via ECAP at RT, and 540 nm grain size was reported in the ECAP-processed alloy.106) It was required to apply a pre-ECAP rolling process to the Zn–0.3Al alloy to prevent crack formation during the ECAP processing of the alloy.107) Rolled + ECAP-processed alloy had a mean grain size of 2000 nm.107) Zn–0.5Cu and Zn–0.8Ag two of the other Zn-based quasi-single phase alloys exhibiting RTS after ECAP. Four passes ECAP that was performed at RT brought about mean grain sizes of 1000 nm and ∼2500 nm for the Zn–0.5Cu108) and Zn–0.8Ag109) alloys, respectively. Besides the Zn-based alloys, Al–30Zn and Mg–8Li alloys were also studied to achieve RTS. Edalati et al.79,80) applied 200 cycles of HPT at RT to the Al–30Zn and Mg–8Li alloys. It was reported that some grain growth occurred during the HPT processing of Al–30Zn, and grain size increased from 210 nm to 280 nm after HPT process.79) It was stated that high HPT cycles accelerated the segregation of Zn atoms at the α/α grain boundaries and resulted in high Zn concentration at these boundaries.78) Significant grain refinement, on the other hand, was achieved in the HPT-processed Mg–8Li alloy in an another study.80) Initial microstructure was completely eliminated and a UFG microstructure having 240 nm grain size was attained after 200 HPT cycles. Furthermore, Li atoms segregated along the Mg-rich α/α boundaries.80)
Room temperature superplastic elongations reported in severely-deformed metals are given in Fig. 5. The highest elongation was achieved in dilute Zn–0.3Al alloy as 1000% at a strain rate of 1 × 10−4 s−1.107) Zn–0.8Ag also exhibited a high elongation of 660% at 1 × 10−4 s−1.109)
Room temperature superplastic elongations of severely-deformed metallic materials.
As also stated above, Zn–Al system is mostly considered for RTS. Microstructural parameters affecting the superplastic elongation in Zn–Al alloys were explained in detail in a previously published overview paper.96) Therefore, a brief summary of these parameters was given in the present study. It was stated that lamellar or elongated grains are not favorable for RTS. 250 nm grain-sized Zn–22Al alloy with some lamellar structure showed a lower elongation to failure of 110% than the sample having 500 nm grain size.101) Similarly, equiaxed grainy morphology was found to be more beneficial to achieve higher superplastic elongation than the elongated one.103) It was concluded that stress concentrations near the lamellar structure and at grain boundaries of elongated grains resulted in crack initiation and caused the premature failure.101,103) Distribution of Zn-rich η and Al-rich α phases in the microstructure was stated to be another important parameter affecting the achieved superplastic elongation in Zn–Al alloys. It was shown η/η and η/α phase boundaries are the most favorable ones for GBS and α/α phase (or grain) boundaries do not slide easily.110–113) Hence, decreasing α/α phase boundaries via eliminating agglomeration of phases in the microstructure promote elongation to failure and result in higher superplastic elongation in Zn–Al alloys. Effect of the fraction of η/η and η/α phase boundaries on the superplasticity of Zn–Al alloy can also be clearly seen from the superplastic elongations of the Zn–22Al, Zn–5Al and Zn–0.3Al. Figure 6 shows the TEM microstructures of these alloys. Mean grain sizes of Zn–22Al, Zn–5Al and Zn–0.3Al were reported to be 200 nm, 540 nm and 2000 nm for the Zn–22Al, Zn–5Al and Zn–0.3Al alloys, respectively.102,106,107) It is clear that almost all of the grain boundaries are in the form of η/η in the Zn–0.3Al. The highest fraction of α/α boundary, on the other hand, exists in Zn–22Al alloy (Fig. 6). Thus, the highest superplastic elongation was achieved in Zn–0.3Al with the largest grain size due to the effective GBS mainly in-between η/η phase boundaries.
Severely-deformed Zn-based alloys exhibit superplasticity at RT corresponding to 0.43–0.45 Tm of the alloys due to their low-melting temperatures.97–109) On the other hand, more recently Edalati et al.79,80) reported RTS in the severely-deformed UFG Al–30Zn and Mg–8Li alloys. These two studies can be considered as the extreme examples of LTS since RT corresponds to only ∼0.35 Tm of the alloys. Furthermore, these works differ from the other RTS-related studies with the main microstructural feature leading to RTS. Edalati et al.79) reported that 480% elongation was achieved in the HPT-processed Al–30Zn alloy while un-processed sample with smaller grain size exhibited quite low elongation lower than 50%. Such a great paradox between the grain sizes and elongation values was explained with the different grain boundary characteristics of the samples. It was stated that 200 turns of HPT did not cause grain refinement but resulted in the segregation of Zn atoms at the α/α grain boundaries. High Zn concentration enhanced the grain-boundary diffusion along these boundaries and made them favorable for GBS. Thus, improved diffusivity at the α/α boundaries activated GBS as the main deformation mechanism and brought about high superplastic elongation even at RT corresponds to 0.36 Tm of Al–30Zn.79) HPT processing of Mg–8Li alloy also resulted in segregation of Li atoms at the α/α (Mg–rich) boundaries and enhanced diffusion capability of these boundaries.80) Such a modification on the grain boundaries resulted in a high superplastic elongation of 440% due to the GBS. These results demonstrate that RTS can be achieved by enhancing grain-boundary diffusion of difficult to slide boundaries with modification of the chemical composition of these boundaries.
It has been well established that GBS is the main deformation mechanism in the conventional superplasticity where the m-value is expected to be ∼0.5.4–6) Similarly, almost all of the studies summarized above stated that LTS was originated from the GBS as the main deformation mechanism. However, it is well known that GBS should be accommodated by some other mechanisms like grain-boundary migration, recrystallization, diffusional flow or dislocation slip in order to prevent cavity formation, particularly at the triple junctions due to the stress concentration.1,2) GBS accommodated by dislocation slip was stated as the main deformation mechanism in some of LTS related studies.29,48,52,57,60,91) In this deformation mechanism dislocations moving along the grain boundaries pile-up at the triple junction of the blocking grain and causes stress concentration. In order to prevent cavity formation, dislocations should slip within the blocking grain and pile-up at the next grain boundary. At the later stage, pile-up dislocations are removed in the grain boundaries as the rate-controlling process of the deformation.1,2) It was shown that such a deformation mechanism leads to an exponent of inverse grain p = 2, the stress exponent of n = 2 and D = Dgb where Dgb is coefficient for grain boundary diffusion in eq. (1).1,2) Based on the achieved values of p = 2 and n = 2 Zherebtsev et al.90) was proposed that slip-accommodated GBS was the main deformation mechanism in the MAF + rolled Ti–6Al–4V alloy. Ko et al.87) also stated that GBS accommodated by matrix dislocation movement were the main deformation mechanism in the ECAP-processed AZ91 and Ti–6Al–4V alloys regarding the low activation energy. Similar results were reported in some other studies too.38,44) Diffusion-accommodated GBS was proposed as an another deformation mechanism in the LTS of severely-deformed metals.27,47) Al-Zubeydi et al.47) demonstrated that calculated activation energy of HPT-processed AZ91 is close to the activation energy of grain boundary diffusion for pure Mg. Regarding the activation energy, m-value and surface appearances of the fractured samples they concluded that GBS accommodated by diffusion was the main deformation mechanisms in AZ91 alloy at low strain rates. Based on the similar observations Lee et al.27) also suggested that diffusion-accommodated GBS was the deformation mechanism in the Al7075 alloy.
Zhang et al.92) proposed that stress concentrations at trigeminal grain boundaries of the FSP-processed Ti–15V–3Cr–3Sn–3Al during the GBS was relaxed with the continuous dynamic recrystallization (CDRX) and phase transformation due to the α-phase precipitation. DRX-accommodated GBS was stated to be main deformation mechanism in the ECAP-processed Mg–10Li–Zn alloy too.56) In an another study, Zhang et al.82) attributed the “necklace” shape formation to the simultaneous GBS and grain rotation. They concluded that when GBS encounters an obstacle grain rotation takes place as an accommodation process and changes grain orientation resulting in the necklace formation.82) Besides the grain rotation, DRX was suggested to be an another accommodation process of GBS in Ref. 82). It was stated that when the group of grains encounter an obstacle during sliding stress concentration and crystal distortion occur at this site. Such a distortion provides the energy required to DRX and brings about the relaxation of stress concentration contributing GBS.82)
It has been well established that a stress exponent of n = 3 and the strain rate sensitivity of m = 0.33 indicate that flow process is controlled by solute drag dislocation glide.2) In the superplastic flow where GBS is the dominant deformation mechanism, on the other hand, the strain rate sensitivity is considered to be m = 0.5 (n = 2).2) However, GBS was stated to be the main deformation mechanism in some severely-deformed alloys exhibiting LTS even though they have strain rate sensitivities ranging between 0.25 and 0.36.25–27,29,35,57,87,91,97,98,102,106,107) Reason of relatively low strain rate sensitive in the presence of GBS was not examined in detail in these studies. However, in previous studies97,114,115) a threshold stress was suggested to be the mean reason for low m-value obtained at low temperatures where GBS is the main deformation mechanism. It has been stated that threshold stress originating from the segregation of impurity atoms at boundaries and their interaction with boundary dislocations must be exceeded so that boundary dislocations escape from the impurity atmosphere and contribute to GBS.97,114,115) Thus, threshold stress is directly related to the test temperature, and decreasing temperature increases the threshold stress due to lowering diffusion coefficient for grain boundary diffusion.97,116) Any increase in the threshold stress increases also the flow stress at low strain rates and brings about low m-value. To provide an evidence to this suggestion it was demonstrated that Zn–22% Al did not exhibit Region-I where strain rate sensitivity takes low values when the level of impurities was decreased down to 6 ppm.117) All these observations show that low m-values were observed in some alloys exhibiting LTS due to the relatively low deformation temperature and thus high threshold stress.
There are several benefits and innovation potentials of LTS considering both the superplastic forming technology and applications of superplastic materials. As stated in the introduction section energy consumption constitutes most of the overall cost of manufacturing processes. Achieving conventional superplasticity at high temperatures and at very low strain rates increases the energy consumption due to the requirement of heating up the material to the forming temperature and keep it at that temperature during the manufacturing process. LTS, on the other hand, makes it possible to perform superplastic forming at considerably lower temperatures than the conventional superplasticity. Thus, lower energy requirement due to the lower forming temperature provides energy saving during the forming process. Furthermore, there are several reports on the hydroforming processes of Al5083118) and Mg–8Li57) alloys exhibiting LTS. These studies reveal that superplastic hydroforming in boiling water would be beneficial to reduce the energy consumption. As explained in Section 3, superplasticity can be achieved even at RT in some classes of materials. Furthermore, RT is not limited with the low-melting Zn-based alloys. Al–30Zn79) and Mg–8Li80) alloys with relatively high melting temperature also exhibit RTS via engineering grain-boundary composition and diffusion. These developments in RTS suggest that superplastic forming technology would be possible without any heating. Besides the energy saving, lower forming temperature provides less damage on forming tools and better surface quality of the formed components. It also prevents severe grain growth, and reduce the cavity formation and solute loss from the surface layer which brings about better post-forming properties.13)
Passive seismic damping application of superplastic Zn–22Al alloy can be considered as a unique application of LTS. Tanaka and co-workers9,119,120) investigated the potential of Zn–22Al alloy for being used in seismic damping applications instead of low-yield point steel. They reported that Zn–22Al alloy can be used as the damper material to dissipate the earthquake energy via undergoing plastic deformation. One of the most important advantage of superplastic Zn–22Al alloy as the damper material over the low-yield point steel is that it does not yield strain hardening during deformation. This means in practice that superplastic Zn–22Al seismic dampers can be used effectively for more than one earthquake contrary to the low-yield point steel dampers that should be changed after each earthquake. Tanaka and co-workers’ studies9,119,120) suggest that superplastic materials exhibiting RTS can be used as multi-use seismic damper materials. However, it should be kept in mind that the superplastic material should be thermally stable at RT to be considered as a damping material.
Achieving superplasticity at high temperatures and at very low strain rates is considered to be the most important disadvantage of superplastic forming as compared to the conventional forming processes. Thus, regarding that energy consumption constitutes most of the overall consumption in manufacturing, any attempt to decrease temperature at which superplasticity achieved is very crucial to enhance the applications of superplastic forming. It has been now well established that grain refinement is the key point for low-temperature superplasticity (LTS). It is well documented that some novel grain refinement techniques based on imposing very high strains to the material via introduction of severe plastic deformation (SPD) provides the capability of decreasing grain sizes down to micron or sub-micron range. The present study overviews the available studies aiming to investigate the LTS in a range of metals and alloys after SPD processing, and some conclusions from the investigations are given below.
