Materials Science and Engineering, A 155 (1992) 147-158 147 Mechanical behavior and interface design of MoSi2-based alloys and composites R. Gibala*, A. K. Ghosh, D. 13. Van Aken, D. J. Srolovitz, A. Basu, H. Chang, D. P. Mason and W. Yang Department of Materials Science and Engineering, The University of Michigan, Ann Arbor, M148109-2136 (USA) Abstract The mechanical behavior of hot pressed MoSi2-based composites containing M05Si3, SiO2, CaO and TiC as reinforcing second phases was investigated in the temperature regime 1000-1300 °C. The effects of strain rate on the flow stress for M05Si~-, SiO2- and CaO-containing composites are presented. Effects of several processing routes and microstructural modifications on the mechanical behavior of MoSi2-M05Si ~ composites are given. Of these four composite additions, M05Si 3 and CaO produce strengthening of MoSi 2 in the temperature range investigated. SiO 2 greatly reduces the strength, consistent with the formation of a glassy phase at interface and interphase boundaries. TiC reduces the flow stress of MoSi 2 in a manner that suggests dislocation pumping into the MoSi 2 matrix. The strain rate effects indicate that dislocation creep (glide and climb) processes operate over the temperature range investigated, with some contribution from diffusional processes at the higher temperatures and lower strain rates. Erbium is found to be very effective in refining the microstructures and in increasing the hardness and fracture properties of MoSi2-MosSi 3 eutectics prepared by arc melting. Initial results on microstructural modeling of the deformation and fracture of MoSi2-based composites are also reported. 1. Introduction There is increasing need for high-strength, oxida- tion-resistant materials for elevated temperature struc- tural applications, particularly in aircraft gas turbines and spacecraft air frames. The silicides of refractory metals such as molybdenum, tungsten, niobium and tantalum have great potential as the matrix materials for new composites with service capabilities at temper- atures above approximately 1200°C. In particular, MoSi 2, which melts at 2030 °C, exhibits excellent high- temperature oxidation resistance because of the forma- tion of a protective silica film and has already been used effectively as the major constituent in super kanthal (an MoSi2-SiO 2 alloy) furnace windings for applications to temperatures as high as 1800 °C. The major problems in the use of MoSi 2 for load-bearing structural applications have been inadequate high tem- perature strength at significant applied stresses above 1200°C and poor fracture toughness below the ductile-to-brittle transition temperature (DBTT) of approximately 1000°C. Above 1000 °C, MoSi 2 yields plastically and deforms by dislocation motion similar to the behavior of metallic materials. The need for composite and alloy additions for strengthening at tem- *Present address: Center for Materials Science, Los Alamos National Laboratory, Los Alamos, NM 87545, USA. peratures above 1200 °C and for similar or alternative additions for toughening at lower temperatures has led to extensive investigation of MoSi 2 in the past few years [1]. The purpose of our research is to investigate promising composite approaches to obtain a more fundamental understanding of the mechanical behavior of MoSi 2 and MoSi 2-based composites at both ambient and elevated temperatures. In the present investigation, we report on the effects of several different single com- posite additions, including M05Si3, SIO2, CaO and TiC, all in particulate form. M05Si 3 was investigated as an elastically hard, strong, brittle second phase which can potentially strengthen MoSi 2 at high temperatures and impart low temperature toughness by crack deflection processes. SiO 2 was investigated in small part because of its use in the manufacture of super kanthal, but more importantly because of the role of silica as an interface and interphase contaminant during the processing of MoSi 2 and MoSi2-based composites, which leads to extensive boundary sliding at elevated temperatures. Both of these additions have lower coefficients of thermal expansion (CTE) than MoSi 2 (see Table 1) and thus are capable of generating tensile residual stresses in the hoop direction in the matrix after fabrication and cooling to room temperature. Residual stresses in the radial direction are compressive in nature and are benign from the standpoint of crack generation. CaO 0921-5093/92/$5.00 © 1992 - Elsevier Sequoia. All rights reserved
`
`GE-1015.001
`
`

`
`148 R. Gibala et al. / Mechanical behavior of MoSi 2:based composites TABLE 1. Coefficients of thermal expansion (CTE) and melting temperatures T m for MoSi 2 and several potential particulate additions Material CTE (°C- ~) @ 1000 °C T m (°C) MoSi 2 8.5 x 10 -6 2030 TiC 7.7 x 10 -6 3250 MosSi 3 6.7 x 10 -6 (@ 500 °C) 2180 HfC 6.25 x 10 -6 3890 HfB 2 5.5 × 10 -6 3250 SiO 2 (vitreous) 0.55 x 10 -6 1710 CaO 13.1 X 10 -6 2570 MgO 15.7 x 10 -6 2800 Ce20 14 × 10 -6 2600 was examined as a more stable addition that might reduce boundary sliding. The CTE of CaO is large compared with that of MoSi2, which allows analysis of effects of thermally induced residual stresses. TiC, with a melting temperature above 3000°C, offers the unique prospect of having higher strength than MoSi 2 at low temperatures but comparable or somewhat lower strength and higher plasticity at temperatures near and above the DBTT. The CTE of TiC is approxi- mately the same as that of MoSi 2. These results serve as experimental data with which we can compare parallel attempts to model the deformation and fracture behavior of composite microstructures and to develop methods by which the controlling microstruc- rural parameters and physical properties can be opti- mized. An initial report of these modeling methods and their results is included. 2. Experimental details 2.1. Fabrication of materials MoS~. For most experiments, the monolithic MoSi2 was prepared from as-received powders screened to -325 mesh obtained from Cerac, Incorporated. The powders were hot pressed at 1700°C under 25-30 MPa pressure for 1-2 h in grafoil-lined graphite dies and an argon atmosphere. The typical grain size (d) of MoSi2 after such treatment was 30/zm. The processing schedule for MoSi 2 was always identical with that used for the MoSi2-based composite being compared. In some selected experiments on MoSi 2-MoSSi 3 eutectics, the monolithic material and eutectics of various volume fractions of MosSi 3 were prepared by arc melting under an argon atmosphere. The densities of the hot-pressed MoSi2 specimens were in the range 94%-96%. Porosity in arc-melted MoSi 2 was larger and depended on the exact solidification conditions. Densities ranged from 70%-85%. In some experiments, post-solidifi- cation hot isostatic pressing was used to reduce the range and total amount of porosity observed in arc- melted materials. MoSi2-MosSi ~ composites. Most of the MoSi2-based composites containing MosSi 3 in the range 15-30 vol.% were made by slurry milling mixtures of the MoSi 2 matrix and molybdenum reinforcement powders ( - 325 mesh) for periods of at least 24 h. The mixtures were hot pressed at 1700 °C at 25 MPa for 2 h. This treatment was sufficient to form the MoSi 2-MosSi 3 eutectic mixture in the equilibrium pro- portions of 15%-30% expected from the phase diagram. In some experiments, the eutectic mixtures were prepared by arc melting MoSi 2 and molybdenum powders followed either by hot pressing or hot isostatic pressing. These were done at 1700 °C at 25-28 MPa for 1.5-2 h. Some arc-melted eutectics were also crushed to produce powder and subsequently hot pressed to produce materials with potentially less silica at interface or interphase boundaries. Erbium was added to some arc-melted eutectics as a potential microstructural refiner and as an additional means of removing oxygen from the melt in experiments based on solidification methods. MoS~-SiO 2 composites. These composites which contained approximately 15 vol.% silica were obtained as a commercial super kanthal. The composites were typically 94% dense. MoS~-CaO composites. The MoSi2-CaO compo- sites were made by ball milling CaO and MoSi 2 powders of - 325 mesh for 24 h and then hot pressing the mixtures at 1700 °C at 25 MPa pressure for 2 h. The CaO powder was obtained from Johnson Matthey, Incorporated. The hot-pressed composites of 20 vol.% CaO were 95% dense. MoS~- TiC composites. MoSi 2-10vol.%TiC compo- sites were prepared by dry blending in a ball mill for 20 h prior to hot pressing at 1700 °C at 30 MPa for 1 h. The TiC was obtained from Johnson Matthey, Incorporated as 2.5-4/zm size powder with the stoichi- ometry of TIC0.95. The hot-pressed composite was approximately 95% dense. The TiC particles were well distributed through the matrix, but because of agglomeration the final particle sizes varied from 5 to 60/zm, with a mean value of 10/~m. 2.2. Microstructural characterization Monolithic and composite microstructures were examined by light optical, scanning and transmission electron microscopies (SEM and TEM). Optical and SEM methods were used to assess grain and second-
`
`GE-1015.002
`
`

`
`R. Gibala et al. / Mechanical behavior of MoSi:-based composites 149 phase structure, shape and distribution before and after deformation. TEM was used to characterize the composite substructures in more detail, including analysis of dislocation substructures and phase boundary structures by high-resolution TEM (HRTEM). Conventional TEM studies were done on a JEOL 2000FX microscope, while HRTEM observa- tions were done on a JEOL 4000EX microscope. 2.3. Mechanical properties Microhardness indentation techniques were used to evaluate the temperature dependence of the strength of monolithic MoSi 2 and the several composites investi- gated, as well as the behavior of some of the composite additions. Hot hardness testing was done at tempera- tures of 25-1300 °C with a Nikon QM-2 hot hardness tester. Diamond Vickers indentations at a load of 1 kgf (9.8 N) were used in most experiments. Most mechanical testing was done in compression at a constant crosshead speed corresponding to an initial strain rate of 10 -4 s-L on specimens of dimensions 6 x 3 x 3 mm ~. All tests were done in air at tempera- tures up to 1300 °C. Yttria was used as a platen lubri- cant to minimize friction. Step-strain rate experiments at 10 ~'- 10 -~ s ~ were also performed on most com- posites at several temperatures in the range 1000- 1300°C to determine stress exponents and activation energies for deformation. Most specimens were cut from the hot-pressed compacts or arc-melted and/or hot isostatically pressed ingots using a diamond saw. Compression samples of MoSi:-SiO2 alloys were cut from 3 mm diameter super kanthal heating elements. 3. Results and discussion 3.1. MoSi, The monolithic MoSi 2 prepared by hot pressing exhibited microstructures and mechanical properties similar to those reported previously by other investiga- tors [1, 2]. The typical microstructure of the as- processed material is given in Fig. 1. Regular equiaxed grain sizes of 25-35 #m and densities of 95 _+ 1% with mainly fine dispersed porosity in the range of a few to several micrometers in diameter were observed. Defor- mation had little or no effect on the microstructure given in Fig. 1 except at the higher temperatures and lowest strain rates examined, where dynamic grain growth occurred during mechanical testing. Disloca- tion substructures and dislocation types were much the same as those reported by Umakoshi et al. [3] and Unal et al. [4]. Specifically, dislocations with Burgers vectors {100), (110) and ½(111) were observed in specimens deformed at 1300°C. At 900°C ½{331) dislocations were found [5]. Fig. 1. Grain structure in polarized light of as-hot-pressed MoSi 2, hot pressed at 1700°C for 1 h under 29.4 MPa. The average grain size is 30/~m. The mechanical properties of the monolithic MoSL and the various MoSi2-based composites are given in Figs. 2-7. The results are presented in the form of compressive flow stress-strain rate data obtained from step-strain rate experiments (Figs. 2-4), selected elevated temperature compressive stress-strain curves (Fig. 5), and hot hardness data for the MoSL and selected reinforcements and composites (Figs. 6 and 7). At 1200°C, the compressive flow stress of MoSL ranges from about 300-340 MPa at a strain rate g of 10 4s ~toabout75MPaatg=10 ~'s ~.Thecorre- sponding stress exponent n varies from 5 to 2-2.5 over the same range of strain rates. At 1100 °C, the com- pressive flow stress is about 420 MPa at g = 10 .4 S 1 and falls to 160 MPa at g= 10 (' s 1. The stress exponent changes from about 8 to 3 over this same range of strain rates [6]. The implication is that disloca- tion creep, with both glide and climb contributions, dominates at these temperatures and strain rates. Glide processes dominate at the lower temperatures and higher strain rates, and diffusional creep may con- tribute in addition to climb at the higher temperatures and lower strain rates. in the stress-strain curves of Fig. 5, MoSi2 exhibits significant apparent work hardening, but we have not yet unambiguously sorted out the relative contributions of plastic flow and cracking in these materials as a function of strain. The temperature dependence of the hardness of MoSi 2 up to 1300°C given in Fig. 6 is much more gradual than that typically reported for the yield strength, probably because microhardness does not sample a sufficient volume of material to be affected as much by the rapidly decreasing viscosity of the glassy silica grain boundary phase with increasing
`
`GE-1015.003
`
`

`
`150
`
`R. Gibala et al.
`
`/ Mechanical behavior ofMbSi_,;based composites
`
`temperature to the same extent as uniaxial compres-
`sion or bend tests of bulk materials. The typical inden-
`tation width along the major axis at 1 kgf is 50-80 pm
`over the temperature range 25-1300 °C compared
`with the typical grain size of 30 ,um.
`
`3.2. MoSi2—Mo5Si3 composites
`The microstructure of
`the hot-pressed MoSi2—
`Mo5Si3 particulate composites is typified by the micro-
`graphs for the 30 vol.% composite given in Fig. 8. The
`MoSi2 matrix in the as-processed state maintains the
`
`E:
`
`1?CD
`F-Y-l
`-U)
`an
`:3
`
`KE
`
`15% Mo5Si3
`Most, 30% Mo5Si3
`
`
`
`3'02
`
`10-7
`
`10-‘
`1o-5
`10*‘
`Strain Rate, s"
`
`10"
`
`Fig. 2. Flow stress as a function of strain rate as determined from
`decremental step-strain rate test in compression conducted on
`MoSi2-based composites at 1 100 °C.
`
`15% M05Si3
`'
`
` si3
`
`\is% Si02
`
`10-7
`
`10-4
`10-5
`10'“
`Strain Rate, s'1
`
`10"
`
`Fig. 3. Flow stress as a function of strain rate as determined from
`decremental step-strain rate test in compression conducted on
`MoSi2-based composites at 1200 °C.
`
`1000
`
`100
`
`STRESS,MPa
`
`HP 1200°
`I-lP+HIP l200°C
`
`HP l300°C
`
`I-IP+l-IIP 1300°C
`
`PM+l-IP l200°C
`
`10
`m4
`
`we
`
`m4
`
`m4
`
`STRAIN RATE, s-1
`
`150 R. Gibala et al. / Mechanical behavior of MoSi2:based composites temperature to the same extent as uniaxial compres- sion or bend tests of bulk materials. The typical inden- tation width along the major axis at 1 kgf is 50-80/zm over the temperature range 25-1300°C compared with the typical grain size of 30/zm. 3.2. MoS~-MosSi 3 composites The microstructure of the hot-pressed MoSi 2- MosSi s particulate composites is typified by the micro- graphs for the 30 vol.% composite given in Fig. 8. The MoSi2 matrix in the as-processed state maintains the ~'100 I 15% Mo$Si 3 ~..........~0% c,o I \ [ MoSi2 ~ 10 10"7 10"6 10"s 10"4 10"3 Strain Rate, s "* Fig. 2. Flow stress as a function of strain rate as determined from decremental step-strain rate test in compression conducted on MoSi2-based composites at 1100 °C. 20% CaO :~ 5~Si3 15% Mo 100 MoSi2~ / /~""15% SiO 2 ..J 10 10-7 10-6 10-5 10-4 10-3 Strain Rate, s "1 Fig. 3. Flow stress as a function of strain rate as determined from decremental step-strain rate test in compression conducted on MoSi2-based composites at 1200 °C. 1000 I [- HP 1200°C~ HP 1300°C / ~// PM+HP 1300°C lOOi f PM+HP 1200°C 10 ........................... 10 -6 10 -5 10 -4 10 -3 STRAIN RATE, s -1 Fig. 4. Compressive stress vs. strain rate test results for MoSi 2- 45vo1.% MosSi3 samples processed by hot pressing arc-melted buttons (HP), hot-pressed and hot isostatically pressed arc- melted buttons (HP+ HIP) and arc-melted buttons which were crushed to produce powder and subsequently hot pressed (PM + HP). Note that the stress exponent is approximately 3 for all processing paths at 1200 °C and 1300 °C. ~e r~ 800 600 400 200 1050 °C -- MoSi2-10 v/oTiC .......... MoSi 2 i i i 0 0 0.10 0.20 0.30 TRUE STRAIN Fig. 5. True stress vs. true strain in compression for MoSi: and MoSi2-10vol.%TiC. The initial strain rate in the constant cross- head speed tests is 10 -4 s- i. 25 r~ 20 < ~ 10 ~ 5 0 i i i i i i 0 200 400 600 800 1000 1200 TEMPERATURE, °C Fig. 6. Vickers hardness vs. temperature for MoSi 2 and TiC tested as separate specimens. ~, 20 + lwt% Er 15 0 500 1000 1500 TEMPERATURE, °C Fig. 7. Hot hardness tests of the erbium-modified and untreated eutectic alloys. Results show an increased hardness for erbium- modified alloys.
`
` S:
`
`5
`
`0
`
`MoSi2-10 v/o TiC
`---------- MOSi2
`
`0
`
`0.10
`
`0.20
`
`0.30
`
`TRUE STRAIN
`
`Fig. 5. True stress vs. true strain in compression for MoSi2 and
`MoSi2—10vol.%TiC. The initial strain rate in the constant cross-
`head speed tests is l0“‘ s‘ ‘.
`
`IOU!
`
`
`
`VICKERSHARDNESS,GPa
`
`N6
`
`)-lUI
`
`I-19
`
`U1 0
`
`0
`
`200 400 600 800 1000 1200
`
`TEMPERATURE, °C
`
`Fig. 6. Vickers hardness vs. temperature for MoSi2 and TiC
`tested as separate specimens.
`
`N O
`
`P-l ca
`
`Mo5Si3-MoSi2+ 1wt% Er
`A l
`
`
`
` I-5 UI VICKERSHARDNESS,GPa
`
`Fig. 4. Compressive stress vs. strain rate test results for MoSi2-
`45vol.%Mo5Si3 samples processed by hot pressing arc-melted
`buttons (HP), hot-pressed and hot
`isostatically pressed arc-
`melted buttons (HP+l-IIP) and arc-melted buttons which were
`crushed to produce powder and subsequently hot pressed
`(PM + HP). Note that the stress exponent is approximately 3 for
`all processing paths at 1200 °C and 1300 °C.
`
`0
`
`500
`
`1000
`
`1500
`
`TEMPERATURE, °C
`
`Fig. 7. Hot hardness tests of the erbium-modified and untreated
`eutectic alloys. Results show an increased hardness for erbium-
`modified alloys.
`
`GE-1015.004
`
`GE-1015.004
`
`

`
`R. Gibala et al. / Mechanical behavior of MoSi_,-based composites 151 Fig. 8. Microstructures of the MoSi2-3Ovol.%MosSi 3 composite (a) before and (b) after deformation at g= 10 -4 s -~ and T = 1200°C. Dynamic grain refinement of the matrix in the vicinity of MosSi 3 is observed in (b) (phase identification, A MoSi2, B Mo~Si3): (a) e=0, d=29/~m; (b) e=0.81, d=25/am and d = 7.5/am. same 30/~m grain size and equiaxed shape given for the monolithic material in Fig. 1. The MosSi 3 exists in relatively large regions and is irregularly distributed in the matrix. Upon deformation at 1200 °C and g = 10 -4 s- 1, the MosSi 3 breaks up into much smaller particles, and the MoSi 2 matrix nearest these particles has' developed into a grain structure of about 8/~m in aver- age grain size. This microstructure constitutes about two thirds of the total area fraction. The remainder of the matrix microstructure consists of larger, slightly elongated grains of MoSi 2 with an average intercept of 25/~m. The mechanical behavior of the hot-pressed MoSi2-M%Si 3 particulate composites at 1100 °C and 1200°C is given in Figs. 2 and 3 respectively, along with data for other hot-pressed particulate composites. The compressive yield stress decreases with a decrease in strain rate and an increase in temperature for all composites. The addition of 15 vol.% MosSi 3 increases the strength of MoSi 2 at all strain rates, but increasing the volume fraction of MosSi 3 to 30% decreases the strength to values below that observed for the mono- lithic material over all strain rates and temperatures investigated. The strengthening of MoSi 2 by 15 vol.% is less at 1200 °C than at 1100 °C, but it is still significant at the lower strain rates. Although the strength of MosSi ~ is greater than that of MoSi 2, the expected strengthening in these compo- sites due to MosSi 3 is not observed. The most probable explanation lies in the dynamic grain refinement, illus- trated in part in Fig. 8, which increases with increasing strain and is especially large for the 30% MosSi 3 com- posite. The presence of smaller grains enhances the rate of dislocation climb and other diffusional creep processes and can produce both the small strengthen- ing effect observed for the 15% MosSi 3 composite and the reduced strength of the 30% MosSi 3 composite. The values of the stress exponent n obtained from Figs. 2 and 3 are consistent with this explanation. The stress exponent for the 15% MosSi 3 composite at 1200°C falls from about 11 at g = 10 .4 S- I to 2.5 at g = 10-6 s ~. At 1100 °C, n falls from about 8 to 5-6 over the same range of strain rates. At 1000 °C, n was in the range 18-11 over these same strain rates. However, n was much smaller for the 30% M%Si 3 composites: 5-2.5 for g= 10 -4 to 10 -6 s -1 at 1200°C and 6-5 for i = 10 -4 to 10 6 s I at 1100 °C [6]. Thus, the overall small values of n, the low flow stresses, and the very substantial dynamic grain refinement of the 30% MosSi 3 composite suggest that the strength of the MoSie-M%Si 3 composites is controlled by grain size effects over a wide range of temperatures and strain rates and is not influenced significantly by other soften- ing processes. The significant strengthening at the lower strain rates for the 15% MosSi 5 composite is
`
`GE-1015.005
`
`

`
`152 R. Gibala et al. / Mechanical behavior of MoSi,-based composites probably associated with the breakdown of the larger particles during deformation. The smaller particles then act to hinder dislocation motion in the manner expected for small particle dispersion strengthening at elevated temperatures. Similar investigations on MoSi2-MosSi 3 composites were carried out on arc-melted materials with micro- structures of the type illustrated in Fig. 9. These materials are near the eutectic composition and contain MosSi 3 as a script microstructure, which can be sub- stantially refined by the addition of small amounts (1 wt.%) of erbium. Microalloying with erbium was also investigated as a means of reducing the oxygen concen- tration in the eutectic material by acting as a gettering agent and removing oxygen from the melt. The presence of rare-earth oxides may also contribute to the creep strength of the eutectic microstructure. The microstructure of Fig. 9 illustrates two effects of erbium additions: the scale of the script microstructure is reduced and the occurrence of pro-eutectic MoSi 2 is greatly diminished. Erbium-rich particles are present in the observed microstructures as inclusions associated with the MosSi 3 phase (B in Fig. 9). Microprobe analy- sis reveals that these particles are rich in oxygen compared with either silicide phase. The volume frac- tion of MosSi3 is larger than that obtained in hot- pressed MoSi2-MosSi 3 composites of Figs. 2, 3 and 8, of order 45% for the materials illustrated in Fig. 9. Mechanical behavior data for the arc-melted MoSi2-MosSi 3 composites are given in Figs. 4 and 7. Figure 4 gives compressive flow stress-strain rate results similar to those presented for the hot-pressed particulate composites in Figs. 2 and 3. Data at 1200°C and 1300°C are presented for arc-melted alloys which were processed by hot pressing (HP), by hot pressing and hot isostatic pressing (HP + HIP), and by crushing to produce powder and subsequent hot pressing (PM + HP). There is no systematic difference in the strength effected by these three processing routes in the results obtained at 1300 °C. At 1200 °C, the HP and HP+HIP materials are consistently stronger than the PM + HP composites. All three sets of materials appear to be characterized by a stress exponent of approximately 3 at both temperatures, consistent with many of those obtained for the hot- pressed particulate materials of Figs. 2 and 3. The strengths appear to be somewhat lower than the hot- pressed particulate materials in Figs. 2 and 3, but con- sistent with possible effects of the larger volume fraction of MosSi 3. We have not yet compared detailed MoSiz matrix microstructures and properties between the hot-pressed and arc-melted materials as a source of these strength differences. The effects of erbium modification on the mechani- cal properties of the arc-melted eutectic composite Fig. 9. Backscattering electron micrographs of the MoSi 2- MosSi3 eutectic: (a) a typical arc-cast eutectic alloy; (b) a I wt.% Er modified eutectic alloy. Note the microstructural refinement with the addition of erbium. An erbium-rich phase appears as bright particles within the MosSi 3 phase (phase identification, A MoSi2, B Mo~Si~).
`
`GE-1015.006
`
`

`
`R. Gibala et al. / Mechanical behavior of MoSi_,-based composites 153 were examined by hot hardness testing methods, as illustrated in Fig. 7. The erbium-treated material was found to have a substantially higher hardness at all temperatures from 25 °C to 1300 °C. This increase in hardness is probably due to the increased constraint placed on the MoSi 2 matrix by the MosSi 3 as the micro- structure becomes more refined. At high temperatures, the increased constraint has its origin in the reduced mean free path of dislocations in the MoSi 2 phase effected by the finer two-phase microstructure. At low temperatures, crack deflection processes are more effective in the finer microstructure. There may also be a residual solid solution strengthening effect and additional strengthening associated with the erbium- based oxide particles present in the MosSi 3 phase. These effects will be sorted out in future experiments. The microstructural refinement resulting from the addition of erbium most likely occurs from a decrease in the interfacial surface energy and a subsequent change in the solidification mechanism. Similar results have been observed in rare-earth-modified AI-Si eutectics [7]. We speculate at this point that the micro- structural refinement results from oxygen gettering by erbium and that removal of these interstitials may improve the fracture characteristics of such MoSi 2- MosSi 3 composite microstructures. Our initial studies to investigate the fracture characteristics have been limited to use of room temperature hardness indenta- tions to observe the extent of cracking and the resulting crack paths in erbium-modified and untreated eutectic alloys. We have found that the crack path is deflected around the finer MosSi 3 particles of the erbium- modified alloys and that these alloys are not as exten- sively cracked as the untreated alloys. Thus, it may be possible to improve the fracture toughness by micro- structural refinement, but we have not yet shown that such improvement is a direct result of microalloying. 3.3. MoSi2-SiO 2 composites The microstructure of the super kanthal used in this study is shown in Fig. 10. It consists of 1-20 ~m discrete particles of SiO2 located at the grain bound- aries of the MoSi 2 matrix which has an average grain size of about 30 /~m. During deformation at 1100-1200°C, the soft SiO2 is squeezed along and between the grain boundaries of the matrix, as illus- trated by arrows in Fig. 10(b). The effect on the mechanical properties is very large. In Figs. 2 and 3, the strength of the MoSi2-SiO 2 composite at 1100 °C and 1200 °C respectively, is well below that of any other composite and the monolithic material. The stress exponent is in the range 2.5-5 at all strain rates investigated at both temperatures and does not change much as a function of strain rate. Fig. 10. Microstructures of the MoSi2-15vol.%SiO 2 composite (a) before and (b) after deformation at g=10 -4 s -1 and T= 1200 °C. The SiO 2 is the dark phase. Intrusion of the SiO 2 phase into the MoSi 2 grain boundaries is observed in (b), e.g. at the arrows. (a) e = 0, d = 20/~m; (b) e = 0.69, d = 18.5/~m.
`
`GE-1015.007
`
`

`
`154 R. Gibala et al. / Mechanical behavior of MoSi2-based composites The significant weakening caused by intentionally added SiO 2 to MoSi 2 is associated with its penetration of the matrix grain boundaries illustrated in Fig. 10. SiO 2 at the grain boundaries acts as a shearable film at the higher temperatures, which eases the sliding of matrix grains. This probably accounts for the MoSi 2 grains remaining equiaxed, even after substantial straining at 1200 °C. The present results on silica- containing MoSi 2 illustrate dramatically the impor- tance of minimizing silica as a grain boundary and phase boundary contaminant during the processing of advanced MoSi2-based composites for applications at temperatures above 1100-1200 °C [8]. 3. 4. MoSi 2- CaO composites The microstructure of hot-pressed MoSi 2 reinforced with 20 vol.% CaO consists of equiaxed grains of MoSi 2 about 23-35/xm in size with smaller 1-15/2m CaO particles distributed regularly throughout the matrix, but primarily along the MoSi2 grain boundaries. On deformation, severe deterioration of the particulate microstructure is observed. The matrix grains deform substantially and become slightly smaller and elon- gated with an aspect ratio of 1.5-2. These observations are illustrated in Fig. 11. The results of mechanical tests for the MoSi2-CaO composites in Figs. 2 and 3 do not disclose negative effects on the strength of MoSi 2 as does SIO2. In fact, over much of the temperature and strain rate ranges investigated, there is a slight improvement in strength. The high CTE value of CaO compared with MoSi2 is expected to lead to separation of the CaO particles from the matrix and grain boundary separation of the matrix, since the calcia is present primarily as a grain boundary phase. The general effect of CaO in both Figs. 2 and 3 appears to be modest strengthening associated with retardation of grain boundary sliding. The values of the stress exponent for the MoSiz-CaO composite were obtained from Figs. 2 and 3 in the range 2-5 at both temperatures (except for the largest strain rates examined at 1100 °C, where n increases with increasing strain rate) [6] and are consistent with such analysis. 3.5. MoSi2- TiC composites TiC has potential as a toughening phase for MoSi 2 in the temperature regime near and above the DBTT of about 1000-1300°C. As a monolithic material, TiC flows plastically to significant strains at these tempera- tures and has a relatively low DBTT of about 600 °C [9]. The plasticity can be further enhanced by decreas- ing the carbon content below the stoichiometry TiC0.97. The density of TiC is relatively low (4.93 g cm-3), and the CTE is nearly the same as that for MoSi 2. More- over, observations to date on MoSiz-TiC composites Fig. 11. Microstructures of the MoSi2-20vol.%CaO composite (a) before and (b) after deformation at g=10 -4 s -l and T= 1200 °C. The CaO particles are the smaller, usually white phase in (a). Extensive plastic deformation of the MoSi 2 grains and fracture of CaO particles are observed in (b). (a) e =0, d = 29/~m; (b) e = 0.69, d = 18 Bm.
`
`GE-1015.008
`
`

`
`R. Gibala et al. / Mechanical behavior of MoSi2-based composites 155 hot pressed at up to 1700 °C or heat treated extensively at 1600°C (e.g. 100 h [10]) have not revealed any reaction layers or other forms of interphase instability. Our current studies have confirmed such findings for the as-processed composite, as well as after subsequent deformation at temperatures in the range 1100- 1200 °C. Figure 12 gives a representative microstructure of the as-processed composite microstructure. TiC appears as the darker phase. The interface remains intact during deformation at 1100-1200 °C, perhaps in part because a Kurdjumov-Sachs-like orientation rela- tionship can exist between the MoSi 2 C11 b structure and the TiC NaC1 type (f.c.c.) structure. This is similar to strong interfaces observed in simpler systems such as b.c.c.-f.c.c, metallic interfaces or B2- )' or B2-~ d interfaces in intermetallic systems [11]. The data given in Fig. 6 illustrate the relative changes in hardness of MoSi2 and TiC as a function of temperature. TiC is much harder than MoSi 2 at room temperature, but softens much more rapidly with increasing temperature. In the temperature range 800-1300 °C, TiC is softer and exhibits more plasticity than MoSi 2. Because of the higher melting temperature of TiC and perhaps because of the effects of silica- induced boundary sliding of most monolithic MoSi 2 materials made to date, TiC becomes harder than MoSi2 at temperatures above 1300 °C. Compression test results on the MoSi2-TiC compo- site