Deformation behavior of Mo5SiB2 at elevated temperatures

Deformation behavior of Mo5SiB2 at elevated temperatures

Materials Science & Engineering A 623 (2015) 124–132 Contents lists available at ScienceDirect Materials Science & Engineering A journal homepage: w...

2MB Sizes 1 Downloads 100 Views

Materials Science & Engineering A 623 (2015) 124–132

Contents lists available at ScienceDirect

Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea

Deformation behavior of Mo5SiB2 at elevated temperatures Kunming Pan a,n, Wei Liu a,b, Laiqi Zhang c, Shizhong Wei a,n, Long You b, Junpin Lin c, Jiwen Li b, Liujie Xu a, Sujuan Zhou a, Mingru Han a a

Henan Engineering Research Center for Wear of Materials, Henan University of Science and Technology, Luoyang, Henan, 471023, China School of Materials Science and Engineering, Henan University of Science and Technology, Luoyang, Henan, 471023, China c State Key Laboratory for Advanced Metals and Materials, University of Science and Technology, Beijing, Beijing, 100083, China b

art ic l e i nf o

a b s t r a c t

Article history: Received 13 August 2014 Received in revised form 7 November 2014 Accepted 11 November 2014 Available online 18 November 2014

During high-temperature compression tests on Mo5SiB2, the deformed microstructure varies with the temperature, strain rate and strain, and their changes have profound effects on the deformation behavior. At 1200 1C, the plastic deformation is provided by dislocation glide on potential slip planes, such as {1 1 0} and {0 0 1}. Due to the limitation of less activated slip systems, the specimens rupture when compressed to the strain of  3% at 1200 1C and 1.67  10  4 s  1. With temperature increasing to 1400 1C, the activation of more independent slip systems increases the deformation capability of the alloy. However, there is an increasing tendency for slip planes to be of an unexpected type (e.g. {1 4 3} and {5 2 3}) as a function of decreasing strain rate and increasing temperature, which is related to dislocation climb. Moreover, dynamic recovery or recrystallization is easy to occur under the conditions of heavy deformation (like 25% strain) and slow strain rate (like 1.67  10  4 s  1), showing the decrease in the dislocation density and flow stress. & 2014 Elsevier B.V. All rights reserved.

Keywords: Mo5SiB2 High-temperature compression Dislocation configuration Glide Climb

1. Introduction The ordered intermetallic Mo5SiB2 is a promising high-temperature structural material because of its high melting temperature, good high-temperature strength, and excellent resistance to oxidation and creep [1–7]. The Mo5SiB2 crystal has a body-centered tetragonal D81 structure (I4/mcm) with lattice parameters of a¼ 6.092 Å and c¼11.067 Å. In a unit cell, 20 Mo, 4 Si and 8 B atoms are situated in layered arrangements along the c axis. This structure has been regarded as the obstacle to plastic deformation or dislocation activity especially at room temperature (RT) [8,9]. However, like most intermetallic compounds, it undergoes a brittle-to-ductile transition (BDT) at high temperatures [10], where the increased ductility is companied by a loss of strength. The feature of high-temperature applications makes it necessary to investigate the deformation mechanisms operative at the temperatures above the BDT temperature. Over the past few years, several studies of the nearly single crystals or Mo5SiB2 constituent of multiphase alloys showed some of possible slip systems to be activated when deformed over a range of temperatures. Meyer and coworkers reported that dislocations with Burgers vectors b parallel to o 1 0 0 4 and

n

Corresponding authors. Tel./fax: þ86 731 64270020. E-mail addresses: [email protected] (K. Pan), [email protected] (S. Wei). http://dx.doi.org/10.1016/j.msea.2014.11.032 0921-5093/& 2014 Elsevier B.V. All rights reserved.

o1 1 0 4 glide on {0 0 1} in Mo5SiB2 in multiphase B-doped Mo5Si3 alloys crept at 1240–1320 1C[3]. The previous work by Ito et al. summarized possible slip systems for Mo5SiB2 and inferred relative ease of slip on [0 0 1]{0 1 0} in light of a hard sphere model of slip [11]. They deformed the [0 2 1] oriented Mo5SiB2 single crystals at 1500 1C and noted unsurprisingly that slip occurred along [0 0 1]{0 1 0}. Hayashi et al. suggested that long straight dislocations with b ¼[0 0 1] were observed to lie on (0 1 0) and tended to align along their edge orientation in the [0 2 1] oriented Mo5SiB2 single crystals crept at 1550 1C[2], similar to the features of dislocation structures observed in compression specimens [12,13]. However, lots of curved dislocations with Burgers vectors b of o1 0 04 , o1 1 0 4 and [0 0 1] were observed in deformation structures in the [0 1 0] oriented specimen. No shear stress applied along these directions indicated that dislocation glide could hardly control the creep behavior of Mo5SiB2 single crystals. o1 0 0 4{0 1 2}, o 0 0 1 4{1 1 0}, o1 1 0 4 {1 1 0}, 1/2o 1 1 1 4{1 1 0} and o0 1 2 4{0 2 1} are the possible slip systems on the basis of the Burgers vectors magnitude and slip plane spacing [11]. All the combinations are physically possible as slip systems and will provide sufficient numbers of independent shears for polycrystalline plasticity. However, for the ordered intermetallic Mo5SiB2, it is unlikely that they all achieve mobility by glide. Since Mo5SiB2 with a complex structure displays a limited range of homogeneity that scarcely changes with temperature increasing, the agglomeration of

K. Pan et al. / Materials Science & Engineering A 623 (2015) 124–132

constitutional vacancies was proposed for a mechanism to explain the appearance of edge dislocations and dislocation loops in annealed Mo5SiB2 alloys [14]. It is conceivable that other mechanisms, such as climb of dislocations, grain boundary slide, vacancy agglomeration and hindrance of second phases, exert influences on plasticity in Mo5SiB2 alloys deformed at elevated temperatures. Therefore, the objective of this work is to examine the compressive behavior and microstructure evolution of Mo5SiB2 alloys at elevated temperatures, and to provide an initial understanding on deformation mechanisms.

2. Experimental procedures 2.1. Mechanical testing and microstructure observation The alloys with a nearly Mo5SiB2 composition were produced from the mixed powders of Mo, Si and B by plasma sparking sintering (SPS) [15]. The relative density measured by the Archimedes method was 99.47% of theoretical density. The microstructure was composed of equiaxed grains ( 1.44 μm) with the Mo5SiB2 phase to the extent of 99.15 vol.%. Compression specimens with dimensions of 2 mm diameter  4 mm height were electro-discharge machined from the Mo5SiB2 alloys. High-temperature compression tests were performed on a mechanical testing machine (YKM-2200) in vacuum (10  3 Pa) at 1400 1C and at 1.67  10  3 to 1.67  10  4 s  1. Due to the limitation of weak deformation capability, in the range of 1100–1200 1C only the slower strain rate of 1.67  10  4 s  1 was adopted to avoid easily failure. The tests were carried out to the different plastic strains in terms of the stage of strain-hardening or strain-softening in the strain-stress curves. The exact plastic strain in each specimen at the end of test was measured by a change in the height. Slices for TEM specimens were cut from deformed specimens, followed by ion milling until perforation. TEM observation was performed on a transition electron microscope (Tecnai, F20) operating at 200 kV.

125

be denoted by the cross product of two vectors P1n and P2n, namely u J P1n  P2n.

3. Results and discussion 3.1. Mechanical testing Fig. 2 shows the stress-strain curves of Mo5SiB2 specimens compressed at elevated temperatures in the range of 1.67  10  3 to 1.67  10  4 s  1, originating from the load-displacement data. The curvilinear shape changes obviously with increasing strain rate and temperature, including the yield stress, tensile strength, ductility and young modulus, but all the curves can be divided into the stages of elastic deformation, hardening or softening. Some of possible slip systems should be activated [10], and their activities depend on strain rates and temperatures to a large extent, as evidenced by the trends of strain-stress curves. Concretely, the rapid strain rates cause an increase in the flow stress and a decrease in the ductility, on the contrary to test temperatures in general. 3.2. Microstructure and property analysis The dislocation microstructures in Mo5SiB2 specimens deformed in the range of 1100–1400 1C at different strain rates have been investigated. The Burgers vectors of dislocations were identified by weak-beam imaging [17] and their line traces were

2.2. Identification of the dislocation line vector u Fig. 1 gives the identification diagram of dislocation line vector u by two projections [16]. The dislocation projects a shadow namely line l1 when the incident light is in a right direction b1. The reciprocal lattice vector P1n can be obtained by drawing a line as a normal vector of the plane which is created by l1 and b1. The relation of two vectors P1n and u can be represented as P1n ? u, since the dislocation line vector u is within this plane. The other diffraction pattern can be achieved by changing the inclined angle. Similarly, the other vector P2n is also perpendicular to u. u can thus

Fig. 1. The identification schematic of the dislocation line vector u by two projections [16].

Fig. 2. The stress-strain curves of Mo5SiB2 specimens compressed at different temperatures in the range of 1.67  10  3-1.67  10  4 s  1.

126

K. Pan et al. / Materials Science & Engineering A 623 (2015) 124–132

determined by crystallographic analysis [14]. The details will be discussed initially in terms of the dislocation microstructures for the specimens deformed at 1400 1C and 1.67  10  3 s  1. The microstructures observed at other rates and temperatures can be then analyzed by the similar methods. It is believed that the dislocation microstructures from compressed specimens are representative of such levels of plastic deformation.

3.2.1. Specimens deformed at 1400 1C Fig. 3 shows typical dislocation microstructures of the specimens deformed to  4% strain at the strain rate of 1.67  10  3 s  1. There exist high density dislocations with an uneven distribution, and most of them are present in tangles and networks. It is apparent that there are three different Burgers vectors about dislocations A, B and C marked in Fig. 3. The dislocation A is out

Fig. 3. Bright-field TEM micrographs used for Burgers vector analysis of dislocations in Mo5SiB2 specimens when compressed to ε  4% at 1400 1C and 1.67  10  3 s  1.

K. Pan et al. / Materials Science & Engineering A 623 (2015) 124–132

127

Fig. 4. Bright-field TEM micrographs of the Mo5SiB2 specimens with the axis of [0 1 0] (a) and the corresponding diffraction patterns (b), as well as the axis of [1 1 0] (c) and the corresponding diffraction patterns (d) when deformed to ε  4% at 1400 1C and 1.67  10  3 s  1. The reciprocal lattice vectors P1* and P2* can be identified by the method described in Fig. 1.

of contrast for the fraction vectors g ¼(4̄ 04), (4  40) and (1  1  2), and then its Burgers vector is given to be b ¼[11̄ 1]/2. The dislocation line trace u can be denoted by the cross product of two reciprocal lattice vectors P1n and P2n from two different injection conditions in Fig. 4, which is consistent with the analysis achieved by the complicated stereographic projection [18]. The line vector of dislocation A can be calculated to be u//[1–1–1], and b  u yields a slip plane of (1 1 0) and the character of the segment is mixed (about 14.91 from edge). The segment B is out of contrast when imaged with diffraction vectors g ¼(4̄ 04̄ ), (11̄ 2), indicating that b is parallel to [1̄ 11]/2. Its line vector can also be deduced to be u//[5–3–3] in terms of crystallographic analysis, and thus it is reasonable to conclude that dislocation B on the plane (01̄ 1) is the characteristics of 231 from screw. For dislocation C, it is mixed in character (26.61 from screw) and moves in the slip system [0 1 0](101̄ ). The glide along the vector b ¼[1̄ 11]/2 products high density of parallel long dislocations B, and hence the cardinal glide direction is the form of [1̄ 11]/2, while a few of shot straight dislocations with the vector b ¼[010] react with them at different angels, leaving lots of dislocation jogs behind. The resulting microstructures can improve the resistance to dislocation motion, as evidence by the appearance of obvious work hardening during the plastic deformation (see Fig. 2). Moreover, a few ring-shaped segments or suspected particles appear in these deformed specimens in Fig. 5(a), and prove to be dislocation loops with b//[1̄ 11] by the STEM image in Fig. 5(b). Note that the dislocation A is slightly curvy and a part of it deviates from the slip direction. In the view of energetics, this

dislocation lies as close to the close-packed direction as possible when both points are fixed by particles or other tree dislocations. Nakao et al. have observed similar curved dislocations (mixed or screw) in Si along the direction [1 1 0] under a mechanical stress [19]. In fact, it is difficult to observe such a structure especially in the materials with bcc lattice structures since the dislocation with both fixed points can be stabilized by dissociation [20–22]. However, a small percent of dislocations in Mo5SiB2 exist as the curvilinear forms instead of dissociation because its stacking fault energy (SFE) is very high. For example, the corresponding SFE is about 890mJ/m2[11], much higher than other metals or alloys (see Table 1,[16]), when the dislocations with b ¼[001] are dissociated to three partial dislocations. This is the reason why it is hard to observe faulted dislocations in deformed Mo5SiB2 specimens, except for one special case that Field et al. [8] found the stacking fault in Mo5SiB2 single crystals after annealing treatment at 1600 1C for 336 h and attributed it to the diffusion, aggregation and oblivion of vacancies just caused by the long time hightemperature annealing. In the deformed specimens o111 4/2 {110} and o010 4{101} are possible slip systems, and hence it is reasonable that these dislocations are present from glide as a preferred choice. Moreover, many nodes originate from the intersection of these dislocations, or are further developed into kinks and jogs (see Fig. 3(a)). These complex dislocation microstructures provide more stop signs for dislocation motion to increase the strength of the material, namely strain-hardening. The dislocation configuration varies significantly when the specimens are compressed to the larger strain of 10% at 1400 1C and 1.67  10  3 s  1, as shown in Fig. 6. The images labeled with

128

K. Pan et al. / Materials Science & Engineering A 623 (2015) 124–132

Fig. 5. (a) The bright-field TEM image�the partial detail of Fig. 2(a), and corresponding diffraction pattern of the deformed Mo5SiB2 specimen confirming the absence of dislocation loops, which are further supported by (b) the STEM image.

Table 1 Stacking fault energies (SFE) of metals and alloys [22], 10  7 mJ/m2. Metals and Alloys

SFE

Metals and Alloys

SFE

Ag Au Cu Cu þ 10%Zn Al Ni

20;16.3 7 1.7 45;32 7 5 75;417 9 35 135 240

Ni-7%Al Ge Si α-Co Zn, Zr, Mg, Ti, Be Stainless steels

90 607 8 517 5 25 250–300 19

Fig. 7. A drawing of partial enlargement for the specimen compressed to ε  10% at 1400 1C and 1.67  10  3 s  1 with 2–20 reflection.

Fig. 6. Bright-field images for Burgers vector analysis of dislocations in the Mo5SiB2 specimens when compressed to ε  10% at 1400 1C and 1.67  10  3 s  1. The images labeled with different diffraction vectors are not shown here.

different diffraction vectors are not shown here. Based on the weak-beam imaging and crystallographic analysis, dislocation segments in Fig. 6 have the following characteristics: A, 251 from edge, [11̄ 1]/2(10-1); B, 40.21 from screw, [111]/2(14̄ 3); C, 10.91 from edge, [111]/2(011̄ ). The dislocation density in the deformed specimen reduces, but the flow stress increases slightly to 710 MPa from 670 MPa at the strain of 4% (Fig. 2(b)), which should be mainly concerned with the dislocation configuration. In Fig. 6, in addition to the long straight dislocations with b ¼[11̄ 1]/2 and many short mixed dislocations with b ¼[111]/2, there appear a set of dislocations in a slip system of [11̄ 1]/2(14̄ 3). The density of nodes and jogs is much bigger than that at the strain of 4%, and their positive effect on the resistance to dislocation motion compensates for the negative aspect caused by a decline in dislocation density. From Fig. 7, a larger version of

Fig. 6, the network of dislocations is caused by the interaction between the active dislocations and forest ones, so that the resistance to plastic deformation increases, as manifested by an increase in the flow stress even when there is a slightly decrease in the dislocation density. It is noted that the plane (14̄ 3) is not close-packed and would not normally be expected to be part of an active slip system. In single crystals oriented in such a way (a non-close-packed direction) as to inhibit the expected slip systems, slip might be activated on some of non-close-packed planes. However, in polycrystalline samples as the grains are not so specifically oriented, it is unlikely that this non-close-packed plane (14̄ 3) is activated prior to the potential glide planes. A reasonable explanation might be that these dislocation configurations are produced from climb (vacancy diffusion) rather than glide [23]. In order to speculate whether the plastic deformation caused by dislocation climb can occur at a sufficiently rate at elevated temperatures, Nabarro et al. [24] has put forward a model which objectively reflects the relationships of dislocation climb rate to multiple parameters such as temperature and strain rate. According to this model, it gives the increase of dislocation climb rate with increasing temperature and strain. The climb mechanism in Mo5SiB2 specimens during compression is being systematically investigated and will be

K. Pan et al. / Materials Science & Engineering A 623 (2015) 124–132

reported. Similarly, the compressive deformation of MoSi2 polycrystalline specimens has been characterized as a function of temperatures by Evans et al. [23]. They suggested that dislocation glide was the deformation mechanism at 1200 1C, while the deformation was controlled by both glide and climb at 1400 1C even in the condition of low strain. It is conceivable that, in addition to glide, the dislocation climb becomes an important mean of deformation when the Mo5SiB2 specimens were compressed at 1400 1C. The high temperature and large strain just promote the occurrence of this process. However, climb weakens the drag force (form dislocation jogs) on the motion of other dislocations, likely accompanied with the dynamic recovery and the disappearance of unlike dislocations by reacting with each other. This is also the reason why the high dislocation density did not depend on a higher level of strain. Therefore, the Mo5SiB2 sample displays a compressive strength value of 300 MPa, apparently lower than the value of 1380 MPa at 1200 1C (see Fig. 2), where the plastic deformation is a glide-controlled process. Theoretically, the climb deformation occurs mainly through atom or vacancy diffusion with the increase of isothermal temperatures and holding times. During deformation, the low strain rate equates with more holding time. In this way, the strain rate should be one of the important factors responsible for affecting the microstructure and mechanical properties of Mo5SiB2 specimens. Fig. 8 shows the TEM images of the specimens compressed to the strain of 4% at 1400 1C and a slower strain rate of 1.67  10  4 s  1. Dislocation density slightly declines compared with the specimen deformed to the same strain at 1.67  10  3 s  1(see Fig. 4(a)). From diffraction contrast and line direction analysis (images with different diffraction vectors are not shown here), these

129

dislocations with three different Burgers vectors were confirmed to have such features: A, 29.61 from screw, [111]/2(101̄ ); B, 46.21 from edge, [11̄ 1]/2(143); C, 14.81 from edge, [011](100). The lower density seems to be attributed to the decrease in the number of [111] dislocations in the deformed microstructure. Moreover, cross-slip (marks “1” and “2” in Fig. 8(a)) makes some [111] dislocations easier bypass nodes or other obstacles, decreasing the dislocation density to a certain extent, and the high stacking fault energy of Mo5SiB2 is beneficial to the occurrence of this process. The cross-slip deformation also proves dislocation B being of screw in character, while reacting with [11̄ 1] dislocations (segments A) gives rise to a small part of dislocation B deviating from its initial line direction, as a mixed dislocation. In addition to the primary slip system [111]/2(101̄ ), the appearance of other nonclose-packed slip planes (like {143}) shows that climb has become one of the important dislocation motion forms under this condition. Fig. 8 (b) is the larger version of Fig. 8(a) from the sample deformed to 4% at 1400 1C and 1.67  10  4 s  1. It is found that some of [111] dislocations and tree ones have a strong interaction, and this manifests itself by lots of nodes, jogs or fragments along the [111] dislocations. Dislocation intersection is likely to produce kinks or jogs with the same Burgers vector as the primary dislocations, as evidenced by their same out-of-contrast conditions (not shown here), while the magnitude is related to the Burgers vector of the tree ones. Typically, jogs have to be three separate classes, because they can get changes under different circumstances [25]. When the jog height is equal to 1–2 atomic spacing, the jogs can move with the active dislocations, leaving behind a row of point defects. When

Fig. 8. (a) Dislocation configuration of the Mo5SiB2 specimen compressed to 4% at 1400 1C and 1.67  10  4 s  1 (images with different diffraction vectors not shown here). (b) A larger version and (c) the corresponding diffraction pattern showing the suspected particles of being dislocation loops.

130

K. Pan et al. / Materials Science & Engineering A 623 (2015) 124–132

Fig. 9. The formation schematic of a dislocation dipole, loop or fragments [26]: (a) the termination of a dipole, (b) the dipole being elongated into a loop, or (c) the loop being broken into small fragments.

Fig. 10. (a) Deformed microstructure with low density dislocations for Mo5SiB2 specimen compressed to 25% at 1400 1C and 1.67  10  4 s  1. (b) Observation on the dislocation array or wall showing the existence of climb deformation during compression.

reaching to the value above 20 nm, both ends of the jog have a weak interaction, and thus they can rotation alone on the glide plane using the jog as a pivot. In between, the dislocation pinned by the jog glides forward in the form of a wavy line, eventually forming a pair of elongated edge-type unlike dislocations, namely a dipole (see Fig. 9). In fact, the edge-type dipoles along the [111] dislocations are observed in the deformed sample, as shown in Fig. 8 (a). Because of a strong stress field between unlike dislocations (the dipole), both dislocations can hardly slip off alone especially in Mo5SiB2 phase with the high stacking fault energy. However, the dipole can be terminated by the ways of cross-slip, climb, jog aggregation and interaction with dislocations [27]. Price et al. [28] has pointed that the dipole termination by climb only occurs at high temperatures. This termination phenomenon has also been observed in the Cu single crystals during the first stage of deformation, where the dipole reacts with a dislocation with the same Burgers vector on the cross-slip plane [29,30]. The above views were presented under certain test conditions, but whether or not the findings completely apply to the Mo5SiB2 phase has yet to be verified. Dislocation climb and intersection are widely found in the formed Mo5SiB2 samples, and should have an effect on the dipole movement. In order to reduce the strain energy, the dipole usually is broken down into an elongated loop, or compressed into lots of small fragments (see Fig. 9). The existences of dipoles, loops and fragments become the obstacles to dislocation motion (see Fig. 8), and this result is represented by the slow decrease in flow stress at 4% (see Fig. 2). If not, there will appear a rapid slump in flow stress at a slower strain rate. The microstructure deformed to 10% is also composed of dislocations with b ¼ o111 4/2 and o011 4, and the density is

Fig. 11. Dislocation configuration of the Mo5SiB2 specimen compressed to 25% at 1400 1C and 1.67  10  4 s  1. Images with different diffraction vectors for Burgers vector analysis are not present here.

slightly lower that after compression to 4% strain (images no shown here). When the strain increases to 25%, the dislocation density is rather low, and the deformed structure consists of distorted or elongated grains interspersed with the cell-shaped structure (arrows in Fig. 10). The formation of this structure should be attributed to the multiplication and interaction of dislocations, especially to climb which has proved to be one of the main deformation mechanisms at this temperature and strain rate. It is conceivable that climb made the edge dislocations align into regular arrays or walls, namely low angle sub-grain boundaries.

K. Pan et al. / Materials Science & Engineering A 623 (2015) 124–132

131

Fig. 12. Bright-field images for Burgers vector analysis of dislocations in the Mo5SiB2 specimens when compressed to (a) ε  1% at 1100 1C and 1.67  10  4 s  1, and (b) to ε  2% at 1200 1C and 1.67  10  4 s  1. The images labeled with different diffraction vectors are not shown here.

Moreover, adjacent fine equiaxed grains (marks “1” and “2”) might form after recrystallization, and the large strain (  25%) just is one of the important hydrodynamic conditions. In a very few area, some dislocations are observed at grain boundaries or near particles in Fig. 11, where the dislocations with the same Burgers vector have such characteristics: 231 from edge, [11̄ 1]/2(52-3). The above results is agreement with the general trend of the strainstress curve in Fig. 2(b), and the deformation process is in the strain-softening stage at this moment. In addition, the appearance of the new non-close-packed plane (52-3) shows that the strain also has an important influence on the climb of dislocations. 3.2.2. Specimens deformed at 1100 1C and 1200 1C For the sample compressed at 1100 1C and 1.67  10  4 s  1, although no obvious variation was observed in the height of specimens before fracture, the TEM analysis shows that there are a set of low density short dislocations parallel to each other, without nodes, jogs or loops in the deformed microstructure (Fig. 12(a)). The plastic deformation of Mo5SiB2 happens with the motion of initial slip system [0 1 0](001), according to the methods of diffraction contrast analysis and dislocation line trace analysis. At 1200 1C and 1.67  10  4 s  1, the Mo5SiB2 samples were compressed to the plastic strain of approximately 3% until they failed. Representative microstructure for the samples compressed to 2% is shown in Fig. 12(b). Overall, the dislocations appear to be present in the forms of independence and nodes with low density. From the same analysis, they are found to consist of the mixed dislocation A in the slip system [11–1]/2(110), edge dislocation B in [010](001), and mixed dislocation C in [101](001). The specimens display the increasing deformation capacity, but are inferior to that deformed at higher temperatures due to the limitation of active slip systems.

4. Conclusions The Mo5SiB2 alloys under compression tests displayed a series of changes in the microstructure and mechanical properties. The specific responses observed are dependent on the temperature, strain rate and strain: (1) For samples compressed at 1100 1C, the BDT behavior of Mo5SiB2 happens with the motion of initial slip system o0104{001}, although there is no obvious variation in the height of specimens before failure.

(2) At 1200 1C, the plastic deformation is produced by dislocation glide on expected slip planes, such as {110}and{001}. Due to the less number of activated independent slip systems, the specimens rupture when compressed to the strain of  3% at 1200 1C and 1.67  10  4 s  1. (3) With temperature increasing to 1400 1C, multiple independent slip systems are activated under an applied stress, and thus the deformation capability of the alloy increases. However, there is an increasing tendency for slip planes to be of an unexpected type (e.g. {143} and {523}) as a function of decreasing strain rate and increasing temperature. These results can be explained on the basis of a significant contribution from the climb of dislocations. Moreover, the rapid strain rate is beneficial to the multiplication of dislocations, while dynamic recovery or recrystallization occurs easily under the conditions of heavy deformation and low rate, manifesting itself in the sharp decrease of dislocation density and flow stress for the specimen compressed to 25% strain at 1400 1C and 1.67  10  4 s  1.

Acknowledgements The work was supported by National Natural Science Foundation of China (No.50871012) and National Key Basic Research Program of China (No.2011CB605502). References [1] F. Wang, A. Shan, X. Dong, J. Wu, T. Nonferr. Metal. Soc 17 (2007) 1242–1246. [2] T. Hayashi, K. Ito, K. Ihara, M. Fujikura, M. Yamaguchi, Intermetallics 12 (2004) 699–704. [3] M. Meyer, Intermetallics 4 (1996) 273–281. [4] J.A.R. Lemberg, Adv. mater 24 (2012) 3445–3480. [5] S.P.J. Kim, J. Phase Equilib. Diff 27 (2006) 605–613. [6] T.A. Parthasarathy, M.G. Mendiratta, D.M. Dimiduk, Acta. Mater. 50 (2002) 1857–1868. [7] L. Yu, Z.K. Li, X. Zheng, J.J. Zhang, H. Liu, R. Bai, H. Wang, Mater. Sci. Eng. A 532 (2012) 392–395. [8] R. Sakidja, J.H. Perepezko, S. Kim, N. Sekido, Acta. Mater. 56 (2008) 5223–5244. [9] L. Zhang, K. Pan, W. Du, M. Wang, J. Lin, X. Ni, H. Shang, J. Sun, Intermetallics 50 (2014) 79–85. [10] L. Zhang, K. Pan, J. Lin, Intermetallics 38 (2013) 49–54. [11] K. Ito, K. Ihara, K. Tanaka, M. Fujikura, M. Yamaguchi, Intermetallics 9 (2001) 591–602. [12] K. Ihara, K. Ito, K. Tanaka, M. Yamaguchi, Mater. Sci. Eng. A 329-331 (2002) 222–227. [13] R. Field, Intermetallics 9 (2001) 863–868. [14] N. Sekido, R. Sakidja, J. Pereperzko, Intermetallics 15 (2007) 1268–1276.

132

K. Pan et al. / Materials Science & Engineering A 623 (2015) 124–132

[15] L. Zhang, K. Pan, J. Wang, J. Lin, Adv. Powder. Technol. 24 (2013) 913–920. [16] X.Y. Huang, The microstructure of materials and its electron microcopy analysis, Metallurgical Industry Press, Beijing, 2008. [17] Y. Ishida, H. Ishida, K. Kohra, H. Ichinose, Philos. Mag. A 42 (1980) 453–462. [18] C.T. Forwood, L.M. Clarebrough, Electron Microscopy of the Interfaces in Metals and Alloys, IOP Publishing Ltd, Bristol, 1991. [19] S. Nakao, T. Ando, M. Shikida, J. Micromech. Microeng 18 (2008) 15026–15032. [20] K. Kang, V.V. Bulatov, W. Cai, P. Natl. Acad. Sci 109 (2012) 15174–15178. [21] L. Wang, H. Bei, T.L. Li, Scripta Mater. 65 (2011) 179–182. [22] V. Vitek, Philos. Mag. A 84 (2004) 415–428.

[23] D.J. Evans, F.J. Scheltens, J.B. Woodhouse, H.L. Fraser, Philos. Mag. A 75 (1997) 1–15. [24] G.W. Groves, A. Kelly, Philos. Mag. A 161 (1969) 977–986. [25] W.C. Dask, Dislocation and mechanical properties of crystals, Wiley, New York, 1957. [26] H. Derek, Introduction to dislocations, Pergamon Press, Oxford, 1975. [27] J. Li, P.R. Swann, Philos. Mag. A 106 (1964) 617–631. [28] P.B. Price, Philos. Mag. A 63 (1961) 449–451. [29] N. Loutat, C.A. Johnson, Philos. Mag. A 84 (1962) 2051–2057. [30] A.S. Tetelman, Philos. Mag. A 83 (1962) 1801–1816.