- Email: [email protected]
Twinning-induced dynamic recrystallization and micro-plastic mechanism during hot-rolling process of a magnesium alloy
Author’s Accepted Manuscript Twinning-induced dynamic recrystallization and micro-plastic mechanism during hot-rolling process of a magnesium alloy Jinhua Peng, Zhen Zhang, Yaozu Li, Wei Zhou, Yucheng Wu www.elsevier.com/locate/msea
PII: DOI: Reference:
S0921-5093(17)30646-9 http://dx.doi.org/10.1016/j.msea.2017.05.037 MSA35053
To appear in: Materials Science & Engineering A Received date: 28 March 2017 Revised date: 7 May 2017 Accepted date: 9 May 2017 Cite this article as: Jinhua Peng, Zhen Zhang, Yaozu Li, Wei Zhou and Yucheng Wu, Twinning-induced dynamic recrystallization and micro-plastic mechanism during hot-rolling process of a magnesium alloy, Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2017.05.037 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Twinning-induced dynamic recrystallization and micro-plastic mechanism during hot-rolling process of a magnesium alloy Jinhua Peng a, Zhen Zhang a,b,*,Yaozu Lia, Wei Zhou a, Yucheng Wu a,b a b
School of Materials Science and Engineering, Hefei University of Technology, Hefei 230009, China
National–Local Joint Research Center of Non-ferrous Metal Materials and Processing Technology, Hefei 230009, China
Abstract: {
̅ } twinning played an important role during hot-rolling process of AZ31
alloy. In the present work, TEM was employed to investigate such twinning related plastic mechanism. The dislocations were usually lying on the basal plane in matrix, while ⃑
⃑ slip was usually activated near grain boundary as well as in twinned area
even when the orientation was quite soft for basal slip. The combination of ⃑ and ⃑
⃑ slip created the characteristic misorientation relationship along 〈
̅ 〉 in twin
region. The activation of non-basal slip met the requirement to accommodate adequate localized stain in twinned area, which made twinning induced DRX kinetically possible. The twin lamella with the highest SF, which was supposed to be the most effective in accommodating imposed strain, showed the highest tendency for localized strain and led to complete DRX inside twinned region. Keywords: Magnesium alloy; Rolling; Twinning; Dynamic recrystallization 1. Introduction Magnesium alloys attained increasing interest during the past few decades due to their prominent performance in weight saving and other outstanding properties such as high specific strength and good castability, etc. In this way they have great potential in many fields especially in automotive and astronomy application. However, magnesium alloys show limited ductility especially when compressed along c-axis. This is also of the greatest industrial interest because of the inevitable development of a characteristic basal texture during wrought processing of magnesium alloys [1-6]. It is widely accepted that, based on the work on magnesium single crystals as early as from 1960s [7-12], with c-axis in compression contraction twinning is always involved even when the temperature is as high as 300℃ [13]. This kind of twinning Corresponding author: Zhen Zhang Tel: +86 18756015986 E-mail: [email protected]
1
usually accommodates highly localized strain inside, which leads to internal ductile failure along twin lamellas at early stage [14-17]. M. Lentz et al. [18] however claimed that it was not the twins themselves but the lack of plastic relaxation in the vicinity of twin domains that caused failure. They provided solid evidence for somewhat easier activation of ⃑
⃑ slip near twin boundary in Mg-4wt%Li alloy,
which was able to sustain continued straining (from 12% to 33%) after prevalent formation of twinning induced structures. However, in pure Mg and conventional Mg alloys, such as AZ31, ZK60, etc., due to the much higher CRSS of ⃑
⃑ slip, the
stress concentration could not be easily relaxed near twin boundaries and twin area usually becomes preferential sites for localized strain [8-11]. Consequently, the orientation within twinned area is usually difficult to obtain with conventional SEM-EBSD technique, which in turn renders direct investigation on the twinning related plastic mechanism most difficult. This may explain why EBSD results for contraction twins were usually of somewhat poor quality and not quite convincing in magnesium alloys. So far, direct investigation on this twinning related plastic mechanism is scant and the underlying interpretation remains under debate. The intent of this study was to explore such twinning related plasticity mechanism in magnesium alloys. TEM micro-texture analysis was combined with conventional dislocation analyses as well as high resolution observation in order to characterize the twinning behavior in magnesium alloy.
2. Material and methods: The material used in this study was AZ31B sheets fabricated by twin-roll cast (TRC) method. The nominal composition was 3wt% Al, 1wt% Zn and balance Mg. The TRC sheets were initially homogenized at 430 ℃ for 2h. Samples with dimensions of 3cm*10cm*6.4mm were cut from the homogenized sheet and subsequently hot-rolled at 375℃ in single pass with a thickness reduction from 10% to 60%. This rolling temperature was chosen to avoid massive precipitation of β-phase (Mg17Al12) [19-21]. Metallographic observation was conducted on the RD-ND plane (containing rolling direction and normal direction) using Leica EC3 optical microscopy. The samples were mechanically polished and etched using a solution of picric-acetic. (0002) pole figure was measured for samples before 2
hot-rolling on a Panalytical X-ray diffractometer in Schulz reflection geometry. The sample preparation followed standard mechanical polishing procedures with a final immersion in a solution of 20% nitric acid, 80% ethanol for about 10 s to remove any residual stress introduced by sample preparation. TEM observation was carried out on a JEOL 2100F equipment operated at 200KV. The specimens were firstly sectioned parallel to the RD-ND plane, grounded to about 100um, then punched into Ф3mm discs and finally double-jet electrolytically polished to perforation with commercial AC-2 electrolyte at 35.4V, -30℃. 3. Experimental results 3.1 Microstructure evolution during hot-rolling process The starting material showed typical basal texture and equiaxed grains with an average size of around 20μm, as shown in Figure 1a and b. After being rolled with a thickness reduction of 20%, thin twin lamellas extensively formed in original grains as indicated by the arrows in Figure1c. Smaller grains of around 3-5μm were clearly seen inside twin regions at higher magnification (inset A-C). This implied that adequate strain might occur within twin areas, which induced localized dynamic recrystallization (DRX) in these regions. With the increase of thickness reduction, more DRX grains began to form near grain boundaries at the consumption of original matrix. This produced a neck-lace like structure at the rolling reduction of 65%, as shown in Figure 1d. Thin twin lamellas could still be distinguished in original matrix, which implied that twinning did play an essential role in hot-rolling process even up to high strain range. In order to reveal the twinning related micro-plastic mechanism, TEM analyses were conducted. 3.2 TEM observation and analyses 3.2.1 Twinning type identification Figure 2 shows the TEM image of two parallel twin lamellas T1, T2 formed in matrix, with a thickness of around 100-120nm. The diffraction pattern taken from the twinned area revealed a [ ̅
]
(
) misorientation with respect to matrix.
Such was close to the theoretical orientation for { (〈
̅ }
{
̅ } secondary twins
̅ 〉-38°). Among all possible twinning planes that contained [ ̅ ̅ ] as η
direction, the only possible planar trace that matched the geometry of T1, T2 twin lamellas was ( ̅ ̅ ) plane (η was the normal direction of the shear plane for 3
corresponding twinning system; it was also the axis about which twinning misorientation was typically created). Although little evidence could be found for any remnant primary twin area, trace analyses did indicate that the twin lamellas should form by a {
̅ }
{
̅ } twinning process.
In another grain, we did find a twin lamella TA, where the evidence for {
̅ }
{
̅ } secondary twinning was retained, as shown in Figure 3. From the
diffraction pattern, it could be verified that the region A and B were formed by {
̅ } secondary twinning in {
̅ } primary twin area-region C. If further strain
were imposed, the secondary twins would easily spread and consume the whole primary twinned region, which would produce a completely retwined region like the case in twin T1 and T2. 3.2.2 Dislocations in matrix In the matrix between T1 and T2 twin lamellas, planar tangled arrays of dislocations were lying parallel on the basal plane, as shown in Figure 2b. Since the most easily activated slip system in magnesium is basal slip, it was quite reasonable to see plenty of dislocations lying on the basal plane. As those dislocations marked the traces of basal plane, a slight lattice rotation of about 6°along 〈
̅ 〉 was revealed
across the twin lamella. Interestingly, dislocations with different morphology were observed near the grain boundary. They were composed of segments lying on different basal planes connected by non-basal segments, as shown in Figure 2d and Figure 4. The Burgers vectors of these dislocations were analyzed under double beam conditions. Since it is well accepted that stable dislocations in HCP metals are limited to only three types ( ⃑, ⃑ and ⃑
⃑) from self-energy considerations[22], 〈
̅ 〉
(
)
zone axis was required to set up the specific double-beam conditions ( ⃗ and 〈
̅ 〉) for unambiguous identification of dislocation types. The dislocations
near the grain boundary could thus be identified as ⃑ type, since they kept contrast with ⃗
〈
〉 and got extinguished when ⃗
〈
̅ 〉 was applied, as shown in
Figure 4(b) and (c). 3.2.3 Twinning induced dynamic recrystallization In the same grain where T1, T2 twins were formed, another equivalent twin T3 was observed. It had the same misorientation of [ ̅ 4
]-(+44.5°) with matrix as T1
and T2, as shown in Figure 5. These parallel ( ̅ ̅ ) twin lamellas T1, T2, T3 intersected with another twin T4, along which a string of small grains were formed. It was reasonable to believe that DRX occurred locally in T4 region, forming those new grains with low density of dislocations inside. The DRX grains were numbered as A-E respectively, from which all the SADPs were recorded. The local micro-texture orientation was calculated and expressed in Euler angles as seen in Figure 5a. Impressively, all of them showed a misorientation along [ ̅
] axis, which was the
η direction of the primary twinning system, as shown in the discrete pole figure (Figure 5b-d). Since new orientations were created in twin region, it was difficult to identify the primary twinning type based on orientation criterion. The shape of twin lamella T4, on the other hand, was in good alignment with the ( ̅
) plane in
matrix. This might confirm that the lamellar structure evolved from a ( ̅
) twin.
Moreover, it was worth noting that T3 was much thicker than T1 and T2; and more dislocations were found inside T3. This meant that T3 was highly probably a more maturely developed twin lamella, where localized strain began to occur preferentially in this area. In order to get a better understanding of the micro-plastic mechanism for twinned area, the dislocations were carefully examined in this region, as shown in Figure 6. 3.2.4 Dislocations within twinned area It could be seen that the dislocations showed irregular shape in twinned area T3 (Figure 6). Some short segments were observed lying on the basal planes, while longer segments were lying close to the prismatic (
̅ ) planes. The Burgers
vectors of these dislocations were also analyzed under double beam conditions. The dislocations were believed to have a Burges vector of ⃑, since they kept sharp in contrast with ⃗
(
) and got extinct with ⃗
〈
̅ 〉.
In the twin lamella TA, where double twinning events occurred, a low angle boundary was found between secondary twinned region A and B. This could be more clearly seen in the high resolution TEM image, taken with η direction as zone axis (Figure 3(b)). The virtual diffraction pattern via Fourier transformation of the image, as well as the traces of the basal planes, showed that the misorientation between A and B region was high as 5.6°, interestingly again along the [ ̅ ̅ ] direction, as shown in Figure 3b. The dislocations lying in the interface were believed to 5
accommodate such lattice rotation. 4. Discussion Considering the typical basal texture of TRC sheets, it was reasonable to believe that contraction twinning played an important role during hot-rolling process to accommodate compression in c-axis. In the present work, twin lamellas started nucleating from early strain stage; and on further straining DRX tended to occur locally both within twin areas and along original grain boundaries. With continuous development of DRX, more grains were formed near original grain boundaries at higher strain stage, which produced a necklace-like structure with the remnant matrix unconsumed and surrounded by small DRX grains. Compression twinning, in this process, acted as stable lamellar structures inside original grains, which most effectively cut matrix into small pieces. This kind of twins was thought to be an effective strengthening mechanism by reducing the mean free path of gliding dislocations in original matrix [18, 23-24]. On the other hand, in the areas between twin lamellas most dislocations were lying on the basal planes. Most interestingly, ⃑
⃑ slip was successfully activated in the vicinity of twin boundaries. The ⃑
⃑
dislocations emitted from a twin boundary would easily intersect with ⃑ type slipping dislocations, which produced ⃑ segments on basal plane as found in Figure 4. M. Lentz et al. [18] reported similar phenomenon in Mg-Li alloy. Based on the results from EPSC simulation and IGMA analyses, they proposed that due to the lower τ ratio in Mg-Li alloys ⃑ and ⃑
⃑ slip would become the dominant plastic
mode (τ ratio--the ratio of stress thresholds for pyramidal to basal slip). Such slip activity played a vital role in plastic relaxation in the vicinity of twin lamellas, which hindered the formation of micro-crack and voids. This rendered continued straining possible even after prevalent presence of 3D twinning networks. Although the activation of ⃑
⃑ slip is uncommon for AZ31 alloy with higher τ ratio in our
present work, it was still considered reasonable since it was the only slip system to accommodate compression in c-axis especially when influenced by plastic compatibility stress near twin boundaries. The point is that it might not be as prevalent in AZ31 alloy as in Mg-Li alloy. Here, in the present work, the plastic mechanism inside twin areas was given more attention. It needs to be mentioned that the occurrence of DRX within twinned 6
areas indicated localized strain in such regions. To investigate such twinning induced ̅ }
DRX, the Schmid factors (SFs) were calculated for all the six equivalent {
twinning systems for matrix in Figure 2 and 5 (Table 1). Since the dimension of work pieces contracted in the normal direction (ND), extended in the rolling direction (RD) and kept constrain in transverse direction (TD) during rolling process, ( ̅
) and
( ̅ ̅ ) twinning were activated to effectively meet the macro-strain requirement. ( ̅
) twin lamella T4, with the highest SFs, thus facilitated higher localized strain,
which would lead to complete DRX as observed in this present work. Such was also expected from SF calculation for basal slip in Table 2. Clearly, the values were found much higher for the twinned areas than the matrix. For simplicity, the orientation of twin T4 was assumed to be [ ̅ ̅ ]- (-44.5°) before complete DRX. Since twinning created a much softer lamellar area inside matrix, basal slip was expected to proceed inevitably and extensively within the twinned area with further strain. However, these dislocations were not frequently observed within twin lamellas because of their remarkable mobility and could easily sweep across the whole twinned volume. Basal slip activity was likely to have rotated the lattice away from its original orientation. A typical basal orientation of (
)〈
̅ 〉 was taken into account
̅ 〉), since the experimental orientation of
(rolling plane//(0002), rolling direction//〈
matrix was quite close to it. The induced lattice rotation could be seen in Figure 7, where the twin orientation was expressed by TD, RD and ND in complete inverse )[ ̅
pole figure. Since (
] slip made no contribution to any extension in RD
or contraction in ND required by rolling process, only ( (
)[ ̅
)[
̅ ] or
] slip would be reasonably activated. Although the SFs for the two
systems were equal, if by any chance, one slip system for an example(
)[ ̅
],
was activated, the crystal would gradually rotate along R1 axis, following path 1 in the inverse pole figure. As the crystal rotation went on, the other ( would be preferred since its SF grew higher than the ongoing system. When (
)[
)[ (
̅ ] system )[ ̅
]
̅ ] slip was activated and dominated, the rotation would
be along R2 axis and follow path 2. Consequently, ND, RD would move back to the Φ=0 line and TD to its initial position. As a combined result from the activation of two basal slip systems, RD, ND moved along the Φ=0 line in a zig-zag path back towards the starting basal texture, meanwhile TD keeps dynamical stable in its initial position which is aligned with [ ̅
] direction. 7
However, if the basal planes inside the twin lamella were not well aligned with the twin lamella, which was the most common situation, basal slip won’t be able to proceed continuously alone due to the impeding effect from twin boundaries. The unexpected existence of ⃑ dislocations in the ‘soft’ twin region would confirm the activity of ⃑
⃑ slip. Probably influenced by plastic compatibility stress [25] from
twin boundaries, ⃑
⃑ slip was able to be activated, even when the orientation was
quite soft for simple basal slip. This non-basal slip activity helped accommodate the compatibility stain from continuous basal slip and made internal strain simply a shear along the twin lamella. Consequently the twin thickness remained nearly constant. Therefore the contribution of ⃑
⃑ slip along (shear component) and in normal
direction (normal component) of twin lamellas was believed to be the most important in accommodating such twinning induced strain. The magnitudes of these components were shown in Table 3. It could be also seen that Type 1 slip showed the largest components in TD/η direction, the dimension of which was of little change during rolling process, so they were less possible than type 2 slip. Apparently ⃑ from the same family in Type 2, for example [ ̅
⃑ slip
] and [ ̅ ̅ ], played different
roles in accommodating the compatibility strain, depending on the magnitude of corresponding components. [ ̅
] slip was more effective in accommodating
normal component while [ ̅ ̅ ], was more effective in shear component. In this way twinning related strain was well accommodated by combination of the two slip systems. The crystal rotation expected from [ ̅
] and [ ̅ ̅ ] slip was also
revealed in Figure 7. The lattice would rotate along R3, R4 axis respectively, following path 3 and 4 with increasing strain. Consequently the orientation inside twinned area would rotate along [ ̅
] direction further away from the starting basal texture,
which was an opposite tack to the situation caused by basal slip. In this way, the combination of ⃑ and ⃑ characteristics along 〈
⃑ slip might cause the impressive misorientation
̅ 〉 axis both for low angle boundaries and DRX grains in
twinned region. Since ⃑
⃑ slip was activated, the disassociation of ⃑
⃑ dislocations
themselves or their intersection with basal slip both had chances to produce immobile ⃑ dislocation segments [26, 27], as shown in Figure 4. Although the dominating mechanism remained unclear, the presence of ⃑ sections could confirm the reaction involving ⃑
⃑ slip activity. It also needs to be stressed that basal slip only provided 8
two independent slip directions in hcp structure, much less than in more symmetrical cubic metals. It is thus reasonable to believe that the activation of ⃑
⃑ slip met the
requirement for the formation of three dimensional substructures in the twinned area, which might provide preferential sites for DRX nucleation and made twinning induced DRX kinetically possible. DRX, in turn, created distinctive substructures locally within twin lamellas. The lattice reorientation consequently introduced rendered localized internal strain possible by cooperative basal slip within these substructures. 5. Summary In summary, during hot rolling process of homogenized AZ31 TRC sheets, twinning related mechanism played an important role. Compression twin lamellas started to nucleate from early strain stage. On further straining, DRX tended to occur locally within twin area and began to initiate at original grain boundaries. With simultaneous occurrence of DRX both along grain boundaries and within twinned area, a necklace-like structure was produced. Thin twin lamellas could still be distinguished in original matrix even up to higher strain range. Moreover in matrix, the dislocations in central part between twin lamellas were always lying on the basal plane while ⃑ dislocations with segments both on basal and non-basal planes were found near grain boundaries. This did imply the activation of ⃑
⃑ slip in matrix.
Similar phenomenon was also reported in Mg-Li alloy, in which the activation thresholds for ⃑
⃑ slip were closer to ⃑ slip [18]. Such was uncommon in AZ31
alloy with higher τ ratio. However it was still considered reasonable since it was the only slip system to accommodate compression in c-axis especially when influenced by plastic compatibility stress near twin boundaries. The activation propensity might not be so prevalent as in Mg-Li alloy. In the present work, the plastic mechanism inside the twin area was given more attention. Mostly interestingly, irregular ⃑ dislocations were observed inside twinning areas even though the orientation was quite easy for basal slip. This implied an unexpected activation of ⃑
⃑ slip, which
could be attributed to the plastic compatibility stress from twin boundaries. The combination of ⃑ and ⃑ relationship along 〈
⃑ slip would create the characteristic misorientation
̅ 〉 in twin region. The activation of non-basal slip met the
requirement for the formation of three dimensional substructures in the twinned area, 9
which provided preferential initiation sites for DRX and made twinning induced DRX kinetically possible. The twin lamella with the highest SF, which was supposed to be the most effective in accommodating imposed strain, showed the highest tendency for localized strain and led to complete DRX inside twin region. Acknowledgement: This work was supported by National Natural Science Funds of China (514010720).
References [1] B. Beausir, S. Suwas, L.S. To´ th, K.W. Neale, J.J. Fundenberger, Acta Mater. 56 (2008) 200. Analysis of texture evolution in magnesium during equal channel angular extrusion [2] B. Beausir, S. Biswas, D.I. Kim, L.S. To´ th, S. Suwas, Acta Mater. 57 (2009) 5061. Analysis of microstructure and texture evolution in pure magnesium during symmetric and asymmetric rolling [3] J.Y. Kang, B. Bacroix, R. Brenner, Scr. Mater. 66 (2012) 654. Evolution of microstructure and texture during planar simple shear of magnesium alloy [4] L. Helis, K. Okayasu, H. Fukutomi, Mater. Sci. Eng. A 430 (2006) 98. Microstructure evolution and texture development during high-temperature uniaxial compression of magnesium alloy AZ31 [5] Jae-Hyung Cho, Hyoung-Wook Kim, Suk-Bong Kang, Tong-Seok Han, Acta Mater. 59 (2011) 5638. Bending behavior, and evolution of texture and microstructure during differential speed warm rolling of AZ31B magnesium alloys [6] Somjeet Biswas, Benoît Beausir, Laszlo S. Toth, Satyam Suwas, Acta Mater. 61 (2013) 5263. Evolution of texture and microstructure during hot torsion of a magnesium alloy [7] E.W. Kelley, W.F. Hosford, Trans. Metall. Soc. AIME 242 (1968) 5. The deformation characteristics of textured magnesium [8] W.H. Hartt, R.E. Reed-Hill, Trans. Metall. Soc. AIME 239 (1967) 1511. The irrational habit of second-order { ̅ } { ̅ } twins in magnesium [9] B.C. Wonsiewicz, W.A. Backofen, Trans. Metall. Soc.AIME 239 (1967) 1422. Plasticity of magnesium crystals [10] W.H. Hartt, R.E. Reed-Hill, Trans. Metall. Soc. AIME 242 (1968) 1127. Internal deformation and fracture of second-order { ̅ } { ̅ } twins in magnesium. [11] E.W. Kelley, W.F. Hosford, Trans. Metall. Soc. AIME 242 (1968) 654. Plane-strain compression of magnesium and magnesium alloy crystals [12] R.E. Reedhill, Trans. Metall. Soc. AIME 218 (1960) 554. A study of the { ̅ } and { ̅ } twinning modes in magnesium [13] Talal Al-Samman, Konstantin D. Molodov, Dmitri A. Molodov, Gunter Gottstein, Satyam Suwas, Acta Mater. 60 (2012) 537. Softening and dynamic recrystallization in magnesium single crystals during c-axis compression. [14] Mark Denis Nave, Matthew Robert Barnett, Scripta Materialia 51 (2004) 881–885. Microstructures and textures of pure magnesium deformed in plane-strain compression 10
[15] M.R. Barnett , Z. Keshavarz, A.G. Beer, X. Ma, Acta Materialia 56 (2008) 5–15. Non-Schmid behaviour during secondary twinning in a polycrystalline magnesium alloy. [16] Q. Ma, H. El Kadiri, A.L. Oppedal, Baird, M.F. Horstemeyera, M. Cherkaoui, Scripta Materialia 64 (2011) 813–816. Twinning and double twinning upon compression of prismatic textures in an AM30 magnesium alloy. [17] M.R. Barnett, Materials Science and Engineering A 25 (2007) 8-16. Twinning and the ductility of magnesium alloys: Part II. “Contraction” twins [18] M. Lentz, M. Risse, N. Schaefer, W. Reimers and I. J. Beyerlein, Strength and ductility with { ̅ } { ̅ } double twinning in a magnesium alloy. Nature Communications 7, Article number: 11068 (2016) doi:10.1038/ncomms11 [19] J.A. del Valle, M.T. Pérez-Prado, O.A. Ruano, Mater. Sci. Eng. A 355 (2003) 68-78. Texture evolution during large-strain hot rolling of the Mg AZ61 alloy [20] T.Al-Samman, G. Gottstein, Scripta Materialia 59 (2008) 760–763. Influence of strain path change on the rolling behavior of twin roll cast magnesium alloy. [21] L. Shanga, I.H. Junga, S. Yuea, R. Vermab, E. Essadiqic, J. Alloys Compd. 492 (2010) 173– 183. An investigation of formation of second phases in microalloyed, AZ31 Mg alloys with Ca, Sr and Ce. [22] A. SEEOER, Thwrie der Bitterfehlatellen, Handbuch der Phyeik, Springer, 1955 [23] Knezevic, M. et al. Deformation twinning in AZ31: Influence on strain hardening and texture evolution, Acta. Mater. 2010(58): 6230–6242. [24] Mu, S., Tang, F. & Gottstein, G. A cluster-type grain interaction deformation texture model accounting for twinning-induced texture and strain-hardening evolution: application to magnesium alloys, Acta Mater. 2014(68): 310–324. [25] J Koike, T Kobayashi, T Mukai, H Watanabe, M Suzuki, K Maruyama, K Higashi, Acta Mater. 51 (2003) 2055. The activity of non-basal slip systems and dynamic recovery at room temperature in fine-grained AZ31B magnesium alloys. [26] M.H. Yoo, Metall. Trans. A 12 (1981) 409. Slip, twinning, and fracture in hexagonal close-packed metals [27] T. Obara, H. Yoshinaga, S. Morozumi, Acta Metall. 21 (1973) 845. { ̅ }〈 ̅ 〉 Slip System in Magnesium
Twining Schmid Factor Twinning system ( ̅ )[ ̅ ̅ ] (̅ )[ ̅ ̅ ] ( ̅ )[ ̅ ̅ ] T1,T2,T3 ( ̅ )[ ̅ ̅ ] T4 (̅ )[ ̅ ̅ ] ( ̅ )[ ̅ ̅ ]
ND(contractio n)
RD(extension)
TD(constrain)
0.21 0.49 0.21 0.49
0.04 0.25 0.17 0.47
0.17 0.24 0.04 0.01
0.38 0.38
-0.05 0.03
0.43 0.34
Table 1 Schmid factors of the six equivalent { 11
̅ } twinning systems for matrix in
Figure 5
Area Matrix T1, T2, T3 T4
Orientation expressed in Euler angles
Schmid Factor for Basal Slip ND RD TD (contraction) (extension) (constrain)
(343.1, 85.2, 105.1) (300.4, 97.0, 104.0)
0.29 0.41
0.27 0.44
0.09 0.12
(28.3, 77.9, 95.8)
0.42
0.39
0.2
Table 2 Schmid factors for basal slip in different areas in Figure 5
Twinning system
Type 1
Type 2
Component in normal Component in shear direction (in unit of direction (in unit of a ) a)
[2113]
1.48
0.66
[2113]
1.48
0.66
[1213]
1.83
0.13
[1213]
1.13
1.45
[1123]
1.83
0.13
[1123]
1.13
1.45
Table 3 The shear and normal component of different a c slip systems
12
Figure 1
c)
a)
A
ND
ND RD
b)
B
B
A
C
RD
50µm
RD
1.0 2.0 3.0 4.0 5.0
50µm
d)
C
D
F
E
TD
E D
F
ND Max=5.5
RD
50µm
Figure 1 (a) The microstructure and (b) (0002) pole figure before rolling, a typical microstructure of contraction twinning in (c)20% and (d)65% rolled sample
Figure 2
Figure 2 a) TEM image of twin lamellas T1, T2 in an original grain; (b) dislocations between twin lamellas; (c) the diffraction pattern between matrix and T2; (d) dislocations near grain boundary
Figure 3
Figure 3 (a) A( 1011) twin formed in matrix; (b) HRTEM image of a Low Angle Boundary (LAB) within twined region
Figure 4
Figure 4 Dislocations in the vicinity of twin lamella T4 under the double beam condition of (a) and (b)
Figure 5
a)
b)
c)
d)
Figure 5 a) The TEM image of twin lamella T3 intersecting with T4, in which DRX was completed; b) , c) , d) pole figure for different areas
Figure 6
Figure 6 Dislocations in the twin lamella T3 under the double beam condition of (a) and (b)
Figure 7
2
1
3
4
2 1
3
4
Figure7 lattice rotation in twinned area expected from
and
slip
PII: DOI: Reference:
S0921-5093(17)30646-9 http://dx.doi.org/10.1016/j.msea.2017.05.037 MSA35053
To appear in: Materials Science & Engineering A Received date: 28 March 2017 Revised date: 7 May 2017 Accepted date: 9 May 2017 Cite this article as: Jinhua Peng, Zhen Zhang, Yaozu Li, Wei Zhou and Yucheng Wu, Twinning-induced dynamic recrystallization and micro-plastic mechanism during hot-rolling process of a magnesium alloy, Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2017.05.037 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Twinning-induced dynamic recrystallization and micro-plastic mechanism during hot-rolling process of a magnesium alloy Jinhua Peng a, Zhen Zhang a,b,*,Yaozu Lia, Wei Zhou a, Yucheng Wu a,b a b
School of Materials Science and Engineering, Hefei University of Technology, Hefei 230009, China
National–Local Joint Research Center of Non-ferrous Metal Materials and Processing Technology, Hefei 230009, China
Abstract: {
̅ } twinning played an important role during hot-rolling process of AZ31
alloy. In the present work, TEM was employed to investigate such twinning related plastic mechanism. The dislocations were usually lying on the basal plane in matrix, while ⃑
⃑ slip was usually activated near grain boundary as well as in twinned area
even when the orientation was quite soft for basal slip. The combination of ⃑ and ⃑
⃑ slip created the characteristic misorientation relationship along 〈
̅ 〉 in twin
region. The activation of non-basal slip met the requirement to accommodate adequate localized stain in twinned area, which made twinning induced DRX kinetically possible. The twin lamella with the highest SF, which was supposed to be the most effective in accommodating imposed strain, showed the highest tendency for localized strain and led to complete DRX inside twinned region. Keywords: Magnesium alloy; Rolling; Twinning; Dynamic recrystallization 1. Introduction Magnesium alloys attained increasing interest during the past few decades due to their prominent performance in weight saving and other outstanding properties such as high specific strength and good castability, etc. In this way they have great potential in many fields especially in automotive and astronomy application. However, magnesium alloys show limited ductility especially when compressed along c-axis. This is also of the greatest industrial interest because of the inevitable development of a characteristic basal texture during wrought processing of magnesium alloys [1-6]. It is widely accepted that, based on the work on magnesium single crystals as early as from 1960s [7-12], with c-axis in compression contraction twinning is always involved even when the temperature is as high as 300℃ [13]. This kind of twinning Corresponding author: Zhen Zhang Tel: +86 18756015986 E-mail: [email protected]
1
usually accommodates highly localized strain inside, which leads to internal ductile failure along twin lamellas at early stage [14-17]. M. Lentz et al. [18] however claimed that it was not the twins themselves but the lack of plastic relaxation in the vicinity of twin domains that caused failure. They provided solid evidence for somewhat easier activation of ⃑
⃑ slip near twin boundary in Mg-4wt%Li alloy,
which was able to sustain continued straining (from 12% to 33%) after prevalent formation of twinning induced structures. However, in pure Mg and conventional Mg alloys, such as AZ31, ZK60, etc., due to the much higher CRSS of ⃑
⃑ slip, the
stress concentration could not be easily relaxed near twin boundaries and twin area usually becomes preferential sites for localized strain [8-11]. Consequently, the orientation within twinned area is usually difficult to obtain with conventional SEM-EBSD technique, which in turn renders direct investigation on the twinning related plastic mechanism most difficult. This may explain why EBSD results for contraction twins were usually of somewhat poor quality and not quite convincing in magnesium alloys. So far, direct investigation on this twinning related plastic mechanism is scant and the underlying interpretation remains under debate. The intent of this study was to explore such twinning related plasticity mechanism in magnesium alloys. TEM micro-texture analysis was combined with conventional dislocation analyses as well as high resolution observation in order to characterize the twinning behavior in magnesium alloy.
2. Material and methods: The material used in this study was AZ31B sheets fabricated by twin-roll cast (TRC) method. The nominal composition was 3wt% Al, 1wt% Zn and balance Mg. The TRC sheets were initially homogenized at 430 ℃ for 2h. Samples with dimensions of 3cm*10cm*6.4mm were cut from the homogenized sheet and subsequently hot-rolled at 375℃ in single pass with a thickness reduction from 10% to 60%. This rolling temperature was chosen to avoid massive precipitation of β-phase (Mg17Al12) [19-21]. Metallographic observation was conducted on the RD-ND plane (containing rolling direction and normal direction) using Leica EC3 optical microscopy. The samples were mechanically polished and etched using a solution of picric-acetic. (0002) pole figure was measured for samples before 2
hot-rolling on a Panalytical X-ray diffractometer in Schulz reflection geometry. The sample preparation followed standard mechanical polishing procedures with a final immersion in a solution of 20% nitric acid, 80% ethanol for about 10 s to remove any residual stress introduced by sample preparation. TEM observation was carried out on a JEOL 2100F equipment operated at 200KV. The specimens were firstly sectioned parallel to the RD-ND plane, grounded to about 100um, then punched into Ф3mm discs and finally double-jet electrolytically polished to perforation with commercial AC-2 electrolyte at 35.4V, -30℃. 3. Experimental results 3.1 Microstructure evolution during hot-rolling process The starting material showed typical basal texture and equiaxed grains with an average size of around 20μm, as shown in Figure 1a and b. After being rolled with a thickness reduction of 20%, thin twin lamellas extensively formed in original grains as indicated by the arrows in Figure1c. Smaller grains of around 3-5μm were clearly seen inside twin regions at higher magnification (inset A-C). This implied that adequate strain might occur within twin areas, which induced localized dynamic recrystallization (DRX) in these regions. With the increase of thickness reduction, more DRX grains began to form near grain boundaries at the consumption of original matrix. This produced a neck-lace like structure at the rolling reduction of 65%, as shown in Figure 1d. Thin twin lamellas could still be distinguished in original matrix, which implied that twinning did play an essential role in hot-rolling process even up to high strain range. In order to reveal the twinning related micro-plastic mechanism, TEM analyses were conducted. 3.2 TEM observation and analyses 3.2.1 Twinning type identification Figure 2 shows the TEM image of two parallel twin lamellas T1, T2 formed in matrix, with a thickness of around 100-120nm. The diffraction pattern taken from the twinned area revealed a [ ̅
]
(
) misorientation with respect to matrix.
Such was close to the theoretical orientation for { (〈
̅ }
{
̅ } secondary twins
̅ 〉-38°). Among all possible twinning planes that contained [ ̅ ̅ ] as η
direction, the only possible planar trace that matched the geometry of T1, T2 twin lamellas was ( ̅ ̅ ) plane (η was the normal direction of the shear plane for 3
corresponding twinning system; it was also the axis about which twinning misorientation was typically created). Although little evidence could be found for any remnant primary twin area, trace analyses did indicate that the twin lamellas should form by a {
̅ }
{
̅ } twinning process.
In another grain, we did find a twin lamella TA, where the evidence for {
̅ }
{
̅ } secondary twinning was retained, as shown in Figure 3. From the
diffraction pattern, it could be verified that the region A and B were formed by {
̅ } secondary twinning in {
̅ } primary twin area-region C. If further strain
were imposed, the secondary twins would easily spread and consume the whole primary twinned region, which would produce a completely retwined region like the case in twin T1 and T2. 3.2.2 Dislocations in matrix In the matrix between T1 and T2 twin lamellas, planar tangled arrays of dislocations were lying parallel on the basal plane, as shown in Figure 2b. Since the most easily activated slip system in magnesium is basal slip, it was quite reasonable to see plenty of dislocations lying on the basal plane. As those dislocations marked the traces of basal plane, a slight lattice rotation of about 6°along 〈
̅ 〉 was revealed
across the twin lamella. Interestingly, dislocations with different morphology were observed near the grain boundary. They were composed of segments lying on different basal planes connected by non-basal segments, as shown in Figure 2d and Figure 4. The Burgers vectors of these dislocations were analyzed under double beam conditions. Since it is well accepted that stable dislocations in HCP metals are limited to only three types ( ⃑, ⃑ and ⃑
⃑) from self-energy considerations[22], 〈
̅ 〉
(
)
zone axis was required to set up the specific double-beam conditions ( ⃗ and 〈
̅ 〉) for unambiguous identification of dislocation types. The dislocations
near the grain boundary could thus be identified as ⃑ type, since they kept contrast with ⃗
〈
〉 and got extinguished when ⃗
〈
̅ 〉 was applied, as shown in
Figure 4(b) and (c). 3.2.3 Twinning induced dynamic recrystallization In the same grain where T1, T2 twins were formed, another equivalent twin T3 was observed. It had the same misorientation of [ ̅ 4
]-(+44.5°) with matrix as T1
and T2, as shown in Figure 5. These parallel ( ̅ ̅ ) twin lamellas T1, T2, T3 intersected with another twin T4, along which a string of small grains were formed. It was reasonable to believe that DRX occurred locally in T4 region, forming those new grains with low density of dislocations inside. The DRX grains were numbered as A-E respectively, from which all the SADPs were recorded. The local micro-texture orientation was calculated and expressed in Euler angles as seen in Figure 5a. Impressively, all of them showed a misorientation along [ ̅
] axis, which was the
η direction of the primary twinning system, as shown in the discrete pole figure (Figure 5b-d). Since new orientations were created in twin region, it was difficult to identify the primary twinning type based on orientation criterion. The shape of twin lamella T4, on the other hand, was in good alignment with the ( ̅
) plane in
matrix. This might confirm that the lamellar structure evolved from a ( ̅
) twin.
Moreover, it was worth noting that T3 was much thicker than T1 and T2; and more dislocations were found inside T3. This meant that T3 was highly probably a more maturely developed twin lamella, where localized strain began to occur preferentially in this area. In order to get a better understanding of the micro-plastic mechanism for twinned area, the dislocations were carefully examined in this region, as shown in Figure 6. 3.2.4 Dislocations within twinned area It could be seen that the dislocations showed irregular shape in twinned area T3 (Figure 6). Some short segments were observed lying on the basal planes, while longer segments were lying close to the prismatic (
̅ ) planes. The Burgers
vectors of these dislocations were also analyzed under double beam conditions. The dislocations were believed to have a Burges vector of ⃑, since they kept sharp in contrast with ⃗
(
) and got extinct with ⃗
〈
̅ 〉.
In the twin lamella TA, where double twinning events occurred, a low angle boundary was found between secondary twinned region A and B. This could be more clearly seen in the high resolution TEM image, taken with η direction as zone axis (Figure 3(b)). The virtual diffraction pattern via Fourier transformation of the image, as well as the traces of the basal planes, showed that the misorientation between A and B region was high as 5.6°, interestingly again along the [ ̅ ̅ ] direction, as shown in Figure 3b. The dislocations lying in the interface were believed to 5
accommodate such lattice rotation. 4. Discussion Considering the typical basal texture of TRC sheets, it was reasonable to believe that contraction twinning played an important role during hot-rolling process to accommodate compression in c-axis. In the present work, twin lamellas started nucleating from early strain stage; and on further straining DRX tended to occur locally both within twin areas and along original grain boundaries. With continuous development of DRX, more grains were formed near original grain boundaries at higher strain stage, which produced a necklace-like structure with the remnant matrix unconsumed and surrounded by small DRX grains. Compression twinning, in this process, acted as stable lamellar structures inside original grains, which most effectively cut matrix into small pieces. This kind of twins was thought to be an effective strengthening mechanism by reducing the mean free path of gliding dislocations in original matrix [18, 23-24]. On the other hand, in the areas between twin lamellas most dislocations were lying on the basal planes. Most interestingly, ⃑
⃑ slip was successfully activated in the vicinity of twin boundaries. The ⃑
⃑
dislocations emitted from a twin boundary would easily intersect with ⃑ type slipping dislocations, which produced ⃑ segments on basal plane as found in Figure 4. M. Lentz et al. [18] reported similar phenomenon in Mg-Li alloy. Based on the results from EPSC simulation and IGMA analyses, they proposed that due to the lower τ ratio in Mg-Li alloys ⃑ and ⃑
⃑ slip would become the dominant plastic
mode (τ ratio--the ratio of stress thresholds for pyramidal to basal slip). Such slip activity played a vital role in plastic relaxation in the vicinity of twin lamellas, which hindered the formation of micro-crack and voids. This rendered continued straining possible even after prevalent presence of 3D twinning networks. Although the activation of ⃑
⃑ slip is uncommon for AZ31 alloy with higher τ ratio in our
present work, it was still considered reasonable since it was the only slip system to accommodate compression in c-axis especially when influenced by plastic compatibility stress near twin boundaries. The point is that it might not be as prevalent in AZ31 alloy as in Mg-Li alloy. Here, in the present work, the plastic mechanism inside twin areas was given more attention. It needs to be mentioned that the occurrence of DRX within twinned 6
areas indicated localized strain in such regions. To investigate such twinning induced ̅ }
DRX, the Schmid factors (SFs) were calculated for all the six equivalent {
twinning systems for matrix in Figure 2 and 5 (Table 1). Since the dimension of work pieces contracted in the normal direction (ND), extended in the rolling direction (RD) and kept constrain in transverse direction (TD) during rolling process, ( ̅
) and
( ̅ ̅ ) twinning were activated to effectively meet the macro-strain requirement. ( ̅
) twin lamella T4, with the highest SFs, thus facilitated higher localized strain,
which would lead to complete DRX as observed in this present work. Such was also expected from SF calculation for basal slip in Table 2. Clearly, the values were found much higher for the twinned areas than the matrix. For simplicity, the orientation of twin T4 was assumed to be [ ̅ ̅ ]- (-44.5°) before complete DRX. Since twinning created a much softer lamellar area inside matrix, basal slip was expected to proceed inevitably and extensively within the twinned area with further strain. However, these dislocations were not frequently observed within twin lamellas because of their remarkable mobility and could easily sweep across the whole twinned volume. Basal slip activity was likely to have rotated the lattice away from its original orientation. A typical basal orientation of (
)〈
̅ 〉 was taken into account
̅ 〉), since the experimental orientation of
(rolling plane//(0002), rolling direction//〈
matrix was quite close to it. The induced lattice rotation could be seen in Figure 7, where the twin orientation was expressed by TD, RD and ND in complete inverse )[ ̅
pole figure. Since (
] slip made no contribution to any extension in RD
or contraction in ND required by rolling process, only ( (
)[ ̅
)[
̅ ] or
] slip would be reasonably activated. Although the SFs for the two
systems were equal, if by any chance, one slip system for an example(
)[ ̅
],
was activated, the crystal would gradually rotate along R1 axis, following path 1 in the inverse pole figure. As the crystal rotation went on, the other ( would be preferred since its SF grew higher than the ongoing system. When (
)[
)[ (
̅ ] system )[ ̅
]
̅ ] slip was activated and dominated, the rotation would
be along R2 axis and follow path 2. Consequently, ND, RD would move back to the Φ=0 line and TD to its initial position. As a combined result from the activation of two basal slip systems, RD, ND moved along the Φ=0 line in a zig-zag path back towards the starting basal texture, meanwhile TD keeps dynamical stable in its initial position which is aligned with [ ̅
] direction. 7
However, if the basal planes inside the twin lamella were not well aligned with the twin lamella, which was the most common situation, basal slip won’t be able to proceed continuously alone due to the impeding effect from twin boundaries. The unexpected existence of ⃑ dislocations in the ‘soft’ twin region would confirm the activity of ⃑
⃑ slip. Probably influenced by plastic compatibility stress [25] from
twin boundaries, ⃑
⃑ slip was able to be activated, even when the orientation was
quite soft for simple basal slip. This non-basal slip activity helped accommodate the compatibility stain from continuous basal slip and made internal strain simply a shear along the twin lamella. Consequently the twin thickness remained nearly constant. Therefore the contribution of ⃑
⃑ slip along (shear component) and in normal
direction (normal component) of twin lamellas was believed to be the most important in accommodating such twinning induced strain. The magnitudes of these components were shown in Table 3. It could be also seen that Type 1 slip showed the largest components in TD/η direction, the dimension of which was of little change during rolling process, so they were less possible than type 2 slip. Apparently ⃑ from the same family in Type 2, for example [ ̅
⃑ slip
] and [ ̅ ̅ ], played different
roles in accommodating the compatibility strain, depending on the magnitude of corresponding components. [ ̅
] slip was more effective in accommodating
normal component while [ ̅ ̅ ], was more effective in shear component. In this way twinning related strain was well accommodated by combination of the two slip systems. The crystal rotation expected from [ ̅
] and [ ̅ ̅ ] slip was also
revealed in Figure 7. The lattice would rotate along R3, R4 axis respectively, following path 3 and 4 with increasing strain. Consequently the orientation inside twinned area would rotate along [ ̅
] direction further away from the starting basal texture,
which was an opposite tack to the situation caused by basal slip. In this way, the combination of ⃑ and ⃑ characteristics along 〈
⃑ slip might cause the impressive misorientation
̅ 〉 axis both for low angle boundaries and DRX grains in
twinned region. Since ⃑
⃑ slip was activated, the disassociation of ⃑
⃑ dislocations
themselves or their intersection with basal slip both had chances to produce immobile ⃑ dislocation segments [26, 27], as shown in Figure 4. Although the dominating mechanism remained unclear, the presence of ⃑ sections could confirm the reaction involving ⃑
⃑ slip activity. It also needs to be stressed that basal slip only provided 8
two independent slip directions in hcp structure, much less than in more symmetrical cubic metals. It is thus reasonable to believe that the activation of ⃑
⃑ slip met the
requirement for the formation of three dimensional substructures in the twinned area, which might provide preferential sites for DRX nucleation and made twinning induced DRX kinetically possible. DRX, in turn, created distinctive substructures locally within twin lamellas. The lattice reorientation consequently introduced rendered localized internal strain possible by cooperative basal slip within these substructures. 5. Summary In summary, during hot rolling process of homogenized AZ31 TRC sheets, twinning related mechanism played an important role. Compression twin lamellas started to nucleate from early strain stage. On further straining, DRX tended to occur locally within twin area and began to initiate at original grain boundaries. With simultaneous occurrence of DRX both along grain boundaries and within twinned area, a necklace-like structure was produced. Thin twin lamellas could still be distinguished in original matrix even up to higher strain range. Moreover in matrix, the dislocations in central part between twin lamellas were always lying on the basal plane while ⃑ dislocations with segments both on basal and non-basal planes were found near grain boundaries. This did imply the activation of ⃑
⃑ slip in matrix.
Similar phenomenon was also reported in Mg-Li alloy, in which the activation thresholds for ⃑
⃑ slip were closer to ⃑ slip [18]. Such was uncommon in AZ31
alloy with higher τ ratio. However it was still considered reasonable since it was the only slip system to accommodate compression in c-axis especially when influenced by plastic compatibility stress near twin boundaries. The activation propensity might not be so prevalent as in Mg-Li alloy. In the present work, the plastic mechanism inside the twin area was given more attention. Mostly interestingly, irregular ⃑ dislocations were observed inside twinning areas even though the orientation was quite easy for basal slip. This implied an unexpected activation of ⃑
⃑ slip, which
could be attributed to the plastic compatibility stress from twin boundaries. The combination of ⃑ and ⃑ relationship along 〈
⃑ slip would create the characteristic misorientation
̅ 〉 in twin region. The activation of non-basal slip met the
requirement for the formation of three dimensional substructures in the twinned area, 9
which provided preferential initiation sites for DRX and made twinning induced DRX kinetically possible. The twin lamella with the highest SF, which was supposed to be the most effective in accommodating imposed strain, showed the highest tendency for localized strain and led to complete DRX inside twin region. Acknowledgement: This work was supported by National Natural Science Funds of China (514010720).
References [1] B. Beausir, S. Suwas, L.S. To´ th, K.W. Neale, J.J. Fundenberger, Acta Mater. 56 (2008) 200. Analysis of texture evolution in magnesium during equal channel angular extrusion [2] B. Beausir, S. Biswas, D.I. Kim, L.S. To´ th, S. Suwas, Acta Mater. 57 (2009) 5061. Analysis of microstructure and texture evolution in pure magnesium during symmetric and asymmetric rolling [3] J.Y. Kang, B. Bacroix, R. Brenner, Scr. Mater. 66 (2012) 654. Evolution of microstructure and texture during planar simple shear of magnesium alloy [4] L. Helis, K. Okayasu, H. Fukutomi, Mater. Sci. Eng. A 430 (2006) 98. Microstructure evolution and texture development during high-temperature uniaxial compression of magnesium alloy AZ31 [5] Jae-Hyung Cho, Hyoung-Wook Kim, Suk-Bong Kang, Tong-Seok Han, Acta Mater. 59 (2011) 5638. Bending behavior, and evolution of texture and microstructure during differential speed warm rolling of AZ31B magnesium alloys [6] Somjeet Biswas, Benoît Beausir, Laszlo S. Toth, Satyam Suwas, Acta Mater. 61 (2013) 5263. Evolution of texture and microstructure during hot torsion of a magnesium alloy [7] E.W. Kelley, W.F. Hosford, Trans. Metall. Soc. AIME 242 (1968) 5. The deformation characteristics of textured magnesium [8] W.H. Hartt, R.E. Reed-Hill, Trans. Metall. Soc. AIME 239 (1967) 1511. The irrational habit of second-order { ̅ } { ̅ } twins in magnesium [9] B.C. Wonsiewicz, W.A. Backofen, Trans. Metall. Soc.AIME 239 (1967) 1422. Plasticity of magnesium crystals [10] W.H. Hartt, R.E. Reed-Hill, Trans. Metall. Soc. AIME 242 (1968) 1127. Internal deformation and fracture of second-order { ̅ } { ̅ } twins in magnesium. [11] E.W. Kelley, W.F. Hosford, Trans. Metall. Soc. AIME 242 (1968) 654. Plane-strain compression of magnesium and magnesium alloy crystals [12] R.E. Reedhill, Trans. Metall. Soc. AIME 218 (1960) 554. A study of the { ̅ } and { ̅ } twinning modes in magnesium [13] Talal Al-Samman, Konstantin D. Molodov, Dmitri A. Molodov, Gunter Gottstein, Satyam Suwas, Acta Mater. 60 (2012) 537. Softening and dynamic recrystallization in magnesium single crystals during c-axis compression. [14] Mark Denis Nave, Matthew Robert Barnett, Scripta Materialia 51 (2004) 881–885. Microstructures and textures of pure magnesium deformed in plane-strain compression 10
[15] M.R. Barnett , Z. Keshavarz, A.G. Beer, X. Ma, Acta Materialia 56 (2008) 5–15. Non-Schmid behaviour during secondary twinning in a polycrystalline magnesium alloy. [16] Q. Ma, H. El Kadiri, A.L. Oppedal, Baird, M.F. Horstemeyera, M. Cherkaoui, Scripta Materialia 64 (2011) 813–816. Twinning and double twinning upon compression of prismatic textures in an AM30 magnesium alloy. [17] M.R. Barnett, Materials Science and Engineering A 25 (2007) 8-16. Twinning and the ductility of magnesium alloys: Part II. “Contraction” twins [18] M. Lentz, M. Risse, N. Schaefer, W. Reimers and I. J. Beyerlein, Strength and ductility with { ̅ } { ̅ } double twinning in a magnesium alloy. Nature Communications 7, Article number: 11068 (2016) doi:10.1038/ncomms11 [19] J.A. del Valle, M.T. Pérez-Prado, O.A. Ruano, Mater. Sci. Eng. A 355 (2003) 68-78. Texture evolution during large-strain hot rolling of the Mg AZ61 alloy [20] T.Al-Samman, G. Gottstein, Scripta Materialia 59 (2008) 760–763. Influence of strain path change on the rolling behavior of twin roll cast magnesium alloy. [21] L. Shanga, I.H. Junga, S. Yuea, R. Vermab, E. Essadiqic, J. Alloys Compd. 492 (2010) 173– 183. An investigation of formation of second phases in microalloyed, AZ31 Mg alloys with Ca, Sr and Ce. [22] A. SEEOER, Thwrie der Bitterfehlatellen, Handbuch der Phyeik, Springer, 1955 [23] Knezevic, M. et al. Deformation twinning in AZ31: Influence on strain hardening and texture evolution, Acta. Mater. 2010(58): 6230–6242. [24] Mu, S., Tang, F. & Gottstein, G. A cluster-type grain interaction deformation texture model accounting for twinning-induced texture and strain-hardening evolution: application to magnesium alloys, Acta Mater. 2014(68): 310–324. [25] J Koike, T Kobayashi, T Mukai, H Watanabe, M Suzuki, K Maruyama, K Higashi, Acta Mater. 51 (2003) 2055. The activity of non-basal slip systems and dynamic recovery at room temperature in fine-grained AZ31B magnesium alloys. [26] M.H. Yoo, Metall. Trans. A 12 (1981) 409. Slip, twinning, and fracture in hexagonal close-packed metals [27] T. Obara, H. Yoshinaga, S. Morozumi, Acta Metall. 21 (1973) 845. { ̅ }〈 ̅ 〉 Slip System in Magnesium
Twining Schmid Factor Twinning system ( ̅ )[ ̅ ̅ ] (̅ )[ ̅ ̅ ] ( ̅ )[ ̅ ̅ ] T1,T2,T3 ( ̅ )[ ̅ ̅ ] T4 (̅ )[ ̅ ̅ ] ( ̅ )[ ̅ ̅ ]
ND(contractio n)
RD(extension)
TD(constrain)
0.21 0.49 0.21 0.49
0.04 0.25 0.17 0.47
0.17 0.24 0.04 0.01
0.38 0.38
-0.05 0.03
0.43 0.34
Table 1 Schmid factors of the six equivalent { 11
̅ } twinning systems for matrix in
Figure 5
Area Matrix T1, T2, T3 T4
Orientation expressed in Euler angles
Schmid Factor for Basal Slip ND RD TD (contraction) (extension) (constrain)
(343.1, 85.2, 105.1) (300.4, 97.0, 104.0)
0.29 0.41
0.27 0.44
0.09 0.12
(28.3, 77.9, 95.8)
0.42
0.39
0.2
Table 2 Schmid factors for basal slip in different areas in Figure 5
Twinning system
Type 1
Type 2
Component in normal Component in shear direction (in unit of direction (in unit of a ) a)
[2113]
1.48
0.66
[2113]
1.48
0.66
[1213]
1.83
0.13
[1213]
1.13
1.45
[1123]
1.83
0.13
[1123]
1.13
1.45
Table 3 The shear and normal component of different a c slip systems
12
Figure 1
c)
a)
A
ND
ND RD
b)
B
B
A
C
RD
50µm
RD
1.0 2.0 3.0 4.0 5.0
50µm
d)
C
D
F
E
TD
E D
F
ND Max=5.5
RD
50µm
Figure 1 (a) The microstructure and (b) (0002) pole figure before rolling, a typical microstructure of contraction twinning in (c)20% and (d)65% rolled sample
Figure 2
Figure 2 a) TEM image of twin lamellas T1, T2 in an original grain; (b) dislocations between twin lamellas; (c) the diffraction pattern between matrix and T2; (d) dislocations near grain boundary
Figure 3
Figure 3 (a) A( 1011) twin formed in matrix; (b) HRTEM image of a Low Angle Boundary (LAB) within twined region
Figure 4
Figure 4 Dislocations in the vicinity of twin lamella T4 under the double beam condition of (a) and (b)
Figure 5
a)
b)
c)
d)
Figure 5 a) The TEM image of twin lamella T3 intersecting with T4, in which DRX was completed; b) , c) , d) pole figure for different areas
Figure 6
Figure 6 Dislocations in the twin lamella T3 under the double beam condition of (a) and (b)
Figure 7
2
1
3
4
2 1
3
4
Figure7 lattice rotation in twinned area expected from
and
slip