The effect of temperature and Fe: Al ratio on the flow and fracture of FeAl

Vol. 43, No. 4, pp. 1723-1730, 1995 Copyright © 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0956-7151/95 $9.50+0.00

Acta metall, mater.

~

Pergamon

0956-7151(94)00369-6

THE EFFECT OF TEMPERATURE A N D Fe:A1 RATIO ON THE FLOW A N D FRACTURE OF FeA1 I. BAKER l, H. XIAO ~, O. KLEIN 1, C. NELSON j and J. D. WHITTENBERGER 2 ~Thayer School of Engineering, Dartmouth College, Hanover, NH 03755 and 2NASA-Lewis Research Center, Brookpark Road, Cleveland, OH 44070, U.S.A.

(Received 16 February 1994,"receivedfor publication 28 September 1994) Abstract--Mechanical testing of large-grained binary FeA1alloys as a function of temperature showed that the yield stress decreased rapidly from 77 to 300 K, followed by a slower decrease up to 500 K. For iron-rich alloys (~<45 at.% AI), the yield stress exhibited a valley at ~ 500 K and a peak at ~ 675 K before declining rapidly at higher temperatures. The stoichiometric alloy showed the highest yield stress at all temperatures. The ductility of the off-stoichiometric alloys in vacuum increased with increasing temperature until at 900 K elongations up to ~ 150% were obtained; the reduction in area showed a sharp transition from < 10% below 500 K to > 80% above 600 K. All off-stoichiometric alloys displayed a mixture of intergranular fracture plus transgranular cleavage below 600 K; the amount of cleavage increased with both decreasing iron concentration and increasing temperature. At temperature I>600 K the fracture mode changed to ductile rupture.

INTRODUCTION FeA1 is a B2 compound which exists over a wide range of composition, from ~ 36.5 to ~ 50 at.% A1 at room temperature [1]. Interestingly, irrespective of composition, the slip vector changes from (111) at room temperature to (100) with increasing temperature [2-6]. (There is disagreement about the compositional dependence of this transition.) Also, cross-slip occurs in Fe-50A1 (compositions given in atomic percent throughout) but slip becomes planar with only a small deviation from the stoichiometric composition [7]. These observations suggest that the mechanical properties vary both with changes in Fe/A1 ratio and with temperature, and experiments have borne out this suggestion. At room temperature, the largest values of both the lattice resistance and the Hall-Petch slope have been measured at the stoichiometric composition [7]. Both these parameters initially decrease rapidly, showing minima around 45 at. % A1, before increasing slightly with increasing iron concentration. The work-hardening rate of as-extruded polycrystalline FeA1 also decreases with increasing Fe concentration [8]. A number of measurements of the mechanical properties of polycrystalline FeA1 as a function of temperature have found that the yield strength either decreases slowly with increasing temperature (fine-grained material) [5, 9, 10] or is independent of temperature (large-grained material) [11-13] from room temperature to some intermediate temperature ( ~ 640 to ~923 K), before decreasing rapidly with further increases in temperature. However, recent studies on both polycrystals and single crystals have shown that a peak in the yield stress exists in the

range 675-873 K for iron-rich FeAI (~<40 at.% A1) [5, 6, 14, 15]. The peak can be obscured in fine-grained material because the grain boundary contribution to the yield strength decreases with increasing temperature, thus, offsetting the increasing lattice strength. Also, reducing the retained vacancies, which raise the room temperature yield strength, enables the yield strength peak to be more clearly revealed [15]. The room temperature fracture of polycrystalline FeAI also depends strongly on the A1 content, with fully intergranular fracture at the stoichiometric composition, becoming first mixed mode and finally wholly transgranular cleavage as the Fe concentration is increased [5, 15-17]. The elongation to fracture increases with increasing iron concentration, from zero at the stoichiometric composition to up to 9% for very iron-rich compositions [16, 17]. Data foriron-rich alloys show an increase in strain to failure from the few percent observed at room temperature to as much as 150% at 900 K [5, 9, 10, 14-18]. In contrast, tensile tests on near-stoichiometric alloys have shown little, if any, elongation to failure at low temperature but at temperatures above 700 K reductions in area of 90% have been reported in [5, 9]. At higher temperatures, both iron-rich and near-stoichiometric alloys have been reported to exhibit grain boundary fracture with evidence of considerable plastic flow and cavitation at the boundaries [10], and dynamic recrystallization was observed in some cases [19]. However, there has also been a report of ductile, dimple-type rupture at high temperature in Fe-40A1 [14]. This paper presents the effect of both temperature and composition on the flow and fracture of

1723

1724

BAKER et al.: THE FLOW AND FRACTURE OF FeAI

polycrystalline FeA1. To observe the lattice properties more clearly, large-grained materials were used, for which the yield strength is approximately the same as the lattice resistance [7]; that is, there was negligible grain boundary contribution to the yield strength. In addition, all specimens were given a long, low-temperature anneal to minimize their vacancy concentration, which reduces their low temperature yield strength.

rate of 2.3 x 10 -4 s -l. Immersion in liquid nitrogen was used for the tests at 77 K. Tests at 173 and 207 K were in methanol cooled by either liquid nitrogen or dry ice. Elevated-temperature compression tests were performed in air on stoichiometric FeA1 using a universal testing machine at an initial strain rate of ~ 1 x 10 -4 s -x. All compression tests were terminated after about 2% strain. The true stress and true strain were calculated with the assumption of constant volume.

EXPERIMENTAL 150 mm long by 38 mm diameter ingots were cast of five FeA1 alloys containing 40, 43, 45, 48 and 50 at.% A1. Rods of the off-stoichiometric alloys were extruded, after vacuum canning in mild steel, at 1173 K at a 16:1 area reduction ratio. A rod was extruded from the canned stoichiometric FeA1 ingot at 1373 K at a 7:1 area reduction ratio. The rods were annealed at temperatures from 1173 to 1473 K to obtain large grain sizes (~230 to ~300/~m), slow cooled at 30 K h -~ and then annealed at 673 K for ~ 1 2 0 h , to remove the retained vacancies [20]. Annealing for an extended period at ~ 673 K after a slow cool, as here, appears to produce the lowest hardness, lowest yield strength and highest ductility for a given Fe:A1 ratio [20,21]. Anneals at temperatures lower than 673 K have little effect on the vacancy concentration in reasonable times (~<40 h) [22]. Dumbbell-shaped tensile specimens (gauge length ~ 18 mm; gauge diameter ~ 3 mm) were ground from the off-stoichiometric rods. In addition, compression specimens (5 mm dia.; 10 mm long, except for Fe-45A1 which were 2.8 nun dia.; 7 mm long) were machined from all the rods. In all cases the long axis of the test specimen was parallel to the extrusion axis. The off-stoichiometric alloys were strained to fracture under tension at temperatures from 298 to 900 K using a universal testing machine at a cross-head speed of 0.017 m m . s -l, corresponding to an initial strain rate of ~1 x 10-4s -1. Tests were performed under a vacuum of ~ 7 x 10 -3 Pa, since it has been shown that the ductility of iron-rich FeAI depends on test environment [23, 24]. The specimens were electro-polished in 10% perchloric acid/90% methanol prior to testing. The strain to fracture was measured directly from the gauge of each fractured specimen. The percentage area reduction after fracture was determined by measuring the cross-sectional area of the fracture surface. Fracture surfaces were examined using a Zeiss DSM 962 scanning electron microscope (SEM) operated at 20 kV. In addition, tensile specimens which had been strained to failure at elevated temperatures were sectioned longitudinally, polished and etched with Marble's reagent. Compression tests were performed at temperatures below room temperature also using a universal testing machine at an initial strain rate of 1.6 x 10 -4 s -l for all but Fe--45AI, for which tests were at an initial strain

RESULTS Yield strength

Yielding occurred smoothly at all temperatures for all alloys, as would be expected in such large-grained material [7]. For the three most iron-rich alloys, Fe--40A1, Fe--43A1 and Fe-45AI, it is evident (see, for example, Fig. 1) that the yield strength (measured as a 0.2% offset stress) does not simply decrease with increasing temperature. Figure 2(a) shows the yield strength as a function of temperature over the entire temperature range for all five alloys; Fig. 2(b), an enlargement of part of Fig. 2(a), shows the yield strength measured under tension as a function of temperature for the off-stoichiometric alloys only. The plots are similar for Fe--40AI, Fe-43A1 and Fe-45A1, viz.; from ~ 80 to 298 K, the yield strength decreases rapidly with increasing temperature; from 298 to ~ 5 0 0 K the yield strength drops slightly with increasing temperature (producing a valley stress); from ~ 500 to 675-700 K the yield strength increases with increasing temperature (producing a peak stress); until, finally, above 675-700 K, the yield strength falls rapidly again. As the aluminum concentration is increased, the difference between the valley stress and the peak stress becomes smaller. Fe--48AI behaves a little differently from the other more iron-rich compositions, see Fig. 2. After the rapid decrease in strength from 77 to 298 K, the yield strength is approximately independent of temperature up to ~ 500 K, above which it falls rapidly again. The behavior of stoichiometric FeA1 is markedly different from the off-stoichiometric alloys, see Fig. 2.

5110 650K 400, ~ 300

G, .201 / / [30K5:

p

..._-- 700K 750K

~ 200 I00

675K

I

2% i

0 TrucSwain Fig. 1. True stress-true strain curve for large-grained Fc-40A1 as a function of temperature. Note that only the initial part of each curve is shown.

BAKER et al.: THE FLOW AND FRACTURE OF FeAI (a) 1250

"~ el

1000

Fe.-40AI

•

Fe-50AI

---~--. Fe43AI -.-.o-.- r~45Aa A r~-4SAt

~.._~_ - ~

~

a

0

, 200

, 400

, 600

Temperature(K)

i 800

10013

(b) 350

]

300 "1

"

~

~.,

---*--. Fe.43AI

•

Fe-40AI

,jr •

.... O"-

Fe-45AI

~5o

0 ) 300

temperature, the initial work hardening rate decreased rapidly. The temperature range over which the work-hardening increased corresponds to the regime of increasing yield stress with increasing temperature, suggesting that the increased work-hardening may be contributing, wholly or partly, to the yield stress increase. Recall that the yield stress is measured not at a sharp yield point but as a 0.2% offset stress.

Fracture of off-stoichiometric alloys

250"

0

1725

, 400

, 500

, 600

, 700

, 800

, 900

1000

Temperature(K) Fig. 2. Graph of yield strength versus temperature for (a) large-grained FeAI of various compositions strained under tension or compression, and (b) large-grained off-stoichiometric FeA1 strained under tension.

Below 500 K fracture occurred after homogeneous deformation, whereas at temperatures above 600 K considerable necking occurred prior to failure. Figure 4(a) shows the elongation to failure of the off-stoichiometric alloys as a function of temperature. All the alloys behaved similarly with the elongation largely increasing with increasing temperature. However, there is a slight drop in elongation at 750-800 K, as noted previously in Fe-37A1-2Ni [5]. At lower temperatures the elongations increased with increasing iron concentration, e.g. from ~ 1.4% in Fe-48AI to ~ 7.4% in Fe-40A1 at room temperature, whereas the high temperature elongation was largely insensitive to the composition with all off-stoichiometric alloys exhibiting large elongations at 900 K. Ductility was also measured as the percentage reduction in area at the tensile specimen's neck. Again,

(a) 12

It is stronger than the other compositions over the temperature range tested, markedly so at low temperature. The decrease in strength between 77 and 298 K is much less than in other alloys, but the small decrease in strength between 298 and 500 K is similar to the most iron-rich alloys. Above 500 K the strength decrease with increasing temperature is dramatic. Interestingly, between 700 and 900 K the yield stress temperature curve parallels the curves for the off-stoichiometric alloys although it is slightly higher.

10

8

g

.=~

6

0

4"

u

o.

2"

, 300

0 20O

, 400

Work-hardening of off-stoichiometric alloys F o r all off-stoichiometric alloys, the work-hardening rates measured at 1% strain decreased with increasing temperature up to ~700 K (see Figs 1 and 3). At temperatures >~75GK after an initial period of work-hardening at low strain, no further work-hardening occurred. At temperatures above 800 K no work-hardening was observed in any alloy. The region of rapidly decreasing yield stress with increasing temperatures roughly coincides with the temperature range where little or no work-hardening was observed. For the three most iron-rich alloys, after an initial decrease with increasing temperature, the workhardening rates measured at 0.2% strain increased with increasing temperature from ~ 500 K to about ~65Oq575 K, see Fig. 3. With further increasing

Oo

, 500

, 600

, 700

800

91)0

700

v S00

900

Temperature (K)

(b) 12"

10'

*~

8'

"~

6'

2'

O

O

u

Ouq S

O

0 200

300

400

500

600

Temperature (K)

Fig. 3. Work-hardening rate at 0.2oVoand 1% strain measured as a function of temperature for (a) Fe-40A1 and (b) Fe-48AI.

BAKER et al.: THE FLOW AND FRACTURE OF FeA1

1726

(a)

500 K; and both Fe-40A1 and Fe--43A1 showed mixed mode fracture at all temperatures ~<500 K, see Fig. 5(a). Some of the lower temperature intergranular fracture surfaces, particularly of Fe-48A1, appeared to be covered with particles, see Fig. 5(b). Examination at higher magnification showed that these 'particles' were the edges of microcleavage facets or cracks, see Fig. 6(b). At temperatures />650 K fracture occurred in a ductile, dimple-type rupture mode in all the off-stoichiometric alloys, Fig. 5(c). At the higher test

160 • •

140

0

120

•

Fe-40AI Fe-43AI Fe45AI F¢-48Al

100

-= ,% 80

!

e• 60 •

4020"

0

O•• •

"t

a 400

200

l

0g o •

600

800

1000

Temperature (K)

(b) 100

~"

80"

o

60.

• • O •

,2Jll e

Fe-40AI Fe-43AI F'c.45AI Fe,-48AI

0

40. < 20'

0 200

0

!

0

I~A A 400

•

600

800

1000

@

Temperature (K)

Fig. 4. Graphs of (a) elongation to failure and (b) reduction in area at the specimen neck as a function of temperature for large-grained FeA1 of various compositions.

the ductility was low at low temperature (-N<500 K), consistent with the region in which homogeneous deformation occurred, see Fig. 4(b). Above 500 K, the reduction in area increased dramatically with increasing temperature to >i 80% at about 600 K. At 900 K, the reduction in area at the neck was in excess of 95 %. Compared with the elongation, the percentage reduction in area increases monotonically with increasing temperature and showed no decrease for any composition. Thus, the elongations obtained largely reflected the length of gauge over which necking occurred. This is the reason for the drop in elongation at 800 K.

Fracture mode of off-stoichiometric alloys Examination of fracture surfaces of the tensile samples showed similar features for all off-stoichiometric alloys. Figure 5 shows some typical micrographs. At temperatures ~<500 K, where the specimens showed low ductility, fracture occurred by either intergranular fracture or a mixture of intergranular fracture and transgranular cleavage. The proportion of transgranular cleavage fracture became greater as the iron content increased and as the temperature increased: Fe--48A1 showed intergranular fracture at all temperatures up to 500 K, see Fig. 5(b); Fe-45A1 showed wholly intergranular fracture up to 400 K but mixed mode fracture between 400 and

,. ii~!~ ¸

Fig. 5. Scanning electron micrographs of fracture surfaces of (a) Fe-40A1 at 500 K, (b) Fe-48AI at 500 K, and (c) Fe-40A1 fractured at 800 K.

BAKER et

al.:

THE FLOW AND FRACTURE OF FeAI

Fig. 6. Scanning electron micrographs of fracture surface of Fe--48A1fractured at 500 K, showing that what appeared to be particles in Fig. 5(b) are in fact microcleavage facets. temperatures (/>750 K) the dimples were so drawn that it was difficult to observe the bottom of the dimples with the SEM. Polished and etched longitudinal sections of the iron-rich alloys again exhibited similar features at a given temperature, see Fig. 7. The grains of specimens tested at 700 K were elongated along the longitudinal direction, indicating that considerable plastic deformation occurs at this temperature, see Fig. 7(a). At 800 K, partial recrystallization occurred at grain boundaries, with greater recrystallization near the fracture surface, see Fig. 7(b). Figure 7(c) is a higher magnification micrograph showing the grain boundary migration that occurs at this temperature. At 900 K, all the material near the fracture surface had recrystallized to a much smaller grain size than the initial grain size. The recrystallization probably occurred after testing since the fracture surface was not decorated with new small grains as previously noted for Fe-37A1-2Ni [19] where dynamic recrystallization was clearly identified.

1727

concentration [25] raises their low temperature yield strength. A rapid decrease in yield strength with increasing temperature below room temperature followed by a slower decrease with further increases in temperature is behavior typical of many B2 compounds [26]. However it is worth noting that a yield strength peak has been noted in several B2 compounds including CoZr [27], CoTi [28] and CuZn [29]. The cause of the strength anomaly in CoZr and CoTi is not clear. Only {110}(001) slip has been observed in CoTi single crystals at temperatures below the peak [28] and the slip systems in CoZr are unknown. The most extensive work has been performed on CuZn for which there are a variety of models that have been proposed to explain the anomalous temperature dependence of the yield strength, including an anomalous elasticity modulus

DISCUSSION A small peak in the yield strength-temperature plot was first observed in a B2-structured FeAl-based alloy by Baker and Gaydosh [5]. Far larger yield strength peaks were subsequently observed in directionally-solidified [14], large-grained polycrystal [15] and single crystal [6] FeAI containing around 40 at.% A1. The present data reconfirm the presence of a yield strength peak and show that the difference between the valley strength and the peak strength decreases with increasing aluminum concentration until for aluminum concentrations of 48 at.% or greater the peak is no longer observed. The peak stress appears to be related to the transition from ( 111 ) to (100) slip that occurs with increasing temperature [2-6]. However, this transition alone does not appear to be able to explain the anomalous temperature dependence since the transition occurs in Fe--48AI and Fe-50AI but no peak was observed in the yield stress. One possible explanation for the lack of observation of a peak in the latter two alloys is that their large vacancy

Fig. 7. Optical micrographs of polished and etched longitudinal sections of fractured specimens of (a) Fe-48A1 at 700 K, (b) Fe-43AI at 800 K, and (c) Fe~5A1 at 800 K.

1728

BAKER et al.: THE FLOW AND FRACTURE OF FeA1

model [29], an antiphase boundary (APB) dragging model [30], a cross-slip model [31] and a climbdissociation model [32-37]. The first of these models depends on the anomalous variation in the (111) shear modulus with temperature [29] whilst the second [30] is based on the idea of having differences in atomic order in the APBs on different planes. Neither of these models can explain the orientation dependence of the critical resolved shear stress (CRSS) [31]. The cross-slip model [31] assumes that (111 ) dislocations cross-slip from their {011} slip plane onto the {112} twinning plane [31]. For the latter model, it has been pointed out that there doesn't appear to be a driving force for the cross-slip process since the anti-phase boundary energy on {112} is higher than that on {110} [38]. The climbdissociation model [32-37] is based on detailed observations of dislocation fine structures using weak-beam transmission electron microscopy. In this work, a correspondence between the yield strength peak temperature and the temperature at which a transition in glide direction occurs from a (111) to a non-a (111) directions was observed. The key feature is that with increasing temperature segments of the (111) dislocations gliding on {110} become climbdissociated. These climbed parts act as a drag on the gliding (111) dislocations. Since the density of climb-dissociated segments increases with increasing temperature, the drag on (111) dislocations and, hence, the yield stress increases with increasing temperature. At the peak temperature, the critical resolved shear stress for (111) slip is so high that other slip vectors become active and a transition occurs in glide direction from (111 ~ to non-(111 ~, specifically, to a mixture of (001) and (110). Whilst this mechanism is attractive, it does not appear to be able to explain the observed strong orientation dependence of the CRSS [38]. Based on earlier observations that (111) dislocations can dissociate into (001) dislocations and (110) dislocations in both CuZn [37, 39] and NiA1 [40] and their own observations that non-(l l l ) slip operates below the peak temperature in FeA1, Yoshima and Hanada [6] proposed that the (111 dislocations in FeAI decompose into (100~ and (110~ dislocations. These decomposition products act as pinning points and increase the CRSS for (111) slip. Since the decomposition is more likely as the temperature increases, more pinning points occur on (111) dislocations until (100) slip becomes easier than (111) slip ((100) dislocations always have the lowest self energy [41]). Above the peak temperature, the yield strength drops rapidly with further increasing temperature because of the decrease in the CRSS of (100) dislocations with increasing in temperature. Their observations are largely in agreement with the earlier work of Baker and Gaydosh [5] who found that in Fe-37A1-2Ni, below the peak temperature most dislocations had a (111) Burgers' vectors but that (100) dislocations were also present, that at the peak

stress temperature most dislocations had <100> Burgers' vector whilst above the peak stress only < 100> dislocations were present. (Some of the ( 100> dislocations present at lower temperatures could have been part of prismatic dislocation loops which arose from vacancy condensation [42-45].) In the experiments reported here, the yield strength peak temperature shows little dependence on aluminum concentration in iron-rich alloys whereas the (111> to (100> slip vector transition with increasing temperature appears to [2-5], although, as noted earlier, there is some disagreement about the compositional dependence of this transition. Thus, any model of the anomalous temperature dependence in FeAI needs to explain this behavior as well as the lack of a yield stress peak in near-stoichiometric or stoichiometric FeA1. A model should also take into account the anomalous temperature dependence of the work-hardening rate, which correlates with the anomalous yield stress behavior. Further, the model should explain why small amounts (few ppm) of boron can shift the yield strength peak to higher temperatures [46]. Further speculation on the origin of the peak seems unwarranted at this point but indicates the need for careful TEM observations, probably on slip-planesectioned strained single crystals. At temperatures ~<500 K, the iron-rich alloys showed low ductility and exhibited either intergranular fracture or intergranular fracture plus transgranular cleavage. The proportion of the latter fracture mode increased with both increasing iron concentration and increasing temperature. As discussed above, within this temperature range (111> slip occurs, which provides sufficient slip systems for plastic deformation in a polycrystal [47]. In addition, TEM in-situ deformation experiments by Baker and Horton [48] demonstrated that dislocation emission is profuse in FeA1. Thus, it is, perhaps, surprising that FeAI is not ductile Liu and George [49] suggested that grain boundaries in FeA1 are intrinsically weak, that is, they lack toughness. The grain boundaries are weakest at the stoichiometric composition, since they fracture at the lowest stress, and become stronger with increasing iron concentration [50]. Liu and George's suggestion is further reinforced by the observations of high quality channeling pattems obtained from intergranular fracture facets of stoichiometric FeAI, which indicate that little local plastic flow occurs during crack propagation in that material [17]. The observation that the stoichiometric material fractures well before yield in tension tests [17, 50] also suggest that dislocations play little part in the fracture of that composition. In contrast, grain boundaries in iron-rich FeAI fracture after plastic flow because the yield stress is below the intrinsic strength of the grain boundaries. In the present study, since transgranular cleavage was also observed, the brittle nature of FeA1 is due not only to the weak grain boundaries but also may be because dislocation cross-slip is difficult. It has

BAKER et al.: THE FLOW AND FRACTURE OF FeAI been suggested that restricted cross-slip in ordered alloys may lead to reduced ductility even if five independent slip systems are available [51, 52]. A mechanism for transgranular cleavage of FeAI based on the formation of (100) dislocation locks has previously been presented [53]. Irrespective of composition, as observed by others [5, 9, 10, 15, 16], the ductility generally increased with increasing temperature. The off-stoichiometric alloys underwent a transition from uniform deformation to necking at temperatures between 500 and 600 K, and at temperatures /> 600 K fracture changed from the low energy brittle fracture modes observed at lower temperatures to more ductile, dimple-type rupture mode. It is worth noting that some previous workers have noted intergranular fracture and grain boundary cavitation at higher temperatures in iron-rich alloys [5,9, 10]. The high temperature grain boundary fracture may be associated with a higher impurity content in those alloys, some of which were produced by powder processing [9]. Although the elongation to failure shows a drop at elevated temperature the reduction in area (RA) at the neck shows only a monotonic increase. That the RA does not show a decrease at any temperature indicates that the elongation not only reflects the ductility of the alloys but also the length of the specimen gauge over which necking occurred. The elongation drop occurs because the work-hardening no longer compensates for the softening due to recovery processes. Thus, necking becomes unstable. The increase in ductility from 500 to 700 K is possibly related to the onset of small amounts o f ( 1 0 0 ) dislocation glide which provide additional deformation modes. It can not be ascribed to the greater mobility of (111) dislocations since their CRSS increases with increasing temperature in this range. At temperatures />700 K (about the peak yield stress temperature in the most iron-rich alloys), even though only (100) slip operates, which provides only three independent slip systems, diffusion-assisted processes such as dislocation climb could have provided the additional deformation modes to allow general plastic deformation. SUMMARY

1729

the peak stress decreases with increasing A1 content, until it disappears in Fe--48A1. 3. F o r the off-stoichiometric alloys, the ductility increases with increasing temperature: between room temperature and 500 K the ductility increase is slight, whilst from 500 to 800 K the elongation more than doubles. At 900 K there is a dramatic increase in elongation (up to ~ 150%). The reduction in area shows a sharp transition from < 10% below 500 K to > 80% above 600 K. 4. For the off-stoichiometric alloys, at low temperature the fracture is mixed-mode fracture, consisting ofintergranular fracture plus transgranular cleavage. The amount of cleavage increases with both increasing iron concentration and increasing temperature. 5. At temperatures >_-600K fracture is by a dimple-type ductile rupture for the off-stoichiometric alloys. This change in fracture mode corresponds to the sharp increases in the reduction in area that occur with increasing temperature. Acknowledgements--This work was supported by the U.S.

Department of Energy, Office of Basic Energy Sciences, Division of Materials Sciences through contract number DE-FG02-87ER45311. The use of the Dartmouth Electron Microscopy Facility is acknowledged. REFERENCES

1. T. B. Massalski, Binary Alloy Phase Diagrams, Vol. 1, p. 112. ASM, Materials Park, Ohio (1986). 2. Y. Umakoshi and M. Yamaguchi, Phil. Mag. A 41, 573 (1980). 3. Y. Umakoshi and M. Yamaguchi, Phil. Mag. A 44, 711 (1981). 4. M.G. Mendiratta, H. K. Kim and H. A. Lipsitt, Metall. Trans. 15A, 395 (1984). 5. I. Baker and D. J. Gaydosh, Mater. Sci. Engng 96, 147 (1987). 6. K. Yoshirni and S. Hanada, in Structural Intermetallics (edited by R. Darolia, J. J. Lewandowski, C. T. Liu, P. L. Martin, D. B. Miracle and M. V. Nathal), p. 475. The Metallurgical Society, Warrendale, Pa. (1993). 7. I. Baker, P. Nagpal, F. Liu and P. R. Munroe, Acta metall, mater. 39, 1637 (1991). 8. B. Schmidt, P. Nagpal and I. Baker, High Temperature Ordered Intermetallics III (edited by C. T. Liu, A. I. Taub, N. S. Stoloff and C. C. Koch), Vol. 133, p. 755. MRS, Pittsburgh, Pa (1989). 9. M. G. Mendiratta, S. K. Ehlers, D. K. Chatterjee and H. A. Lipsitt, in Proc. 3rd Int. Conf. on Rapid Solidification Processing: Materials and Technologies

A combination of tensile and compressive testing of large-grained, low-temperature-annealed FeA1 alloys as a function of temperature (in vacuum at room temperature and above) has shown that:

(edited by R. Mehrabian), p. 240. NBS, Gaithersburg, Md (1983). 10. I. Baker and D. J. Gaydosh, High Temperature Ordered

1. Irrespective of composition, the yield stress decreases rapidly from 77 K to room temperature followed by a slower decrease up to 500 K. The stoichiometric alloy has a much higher yield stress than the other alloys at all test temperatures. 2. For iron-rich alloys (40, 43, 45 at.% AI), the yield stress exhibits a valley at ~ 500 K and a peak at ~ 675 K. The difference between the valley stress and

193 (1986). 12. G. Sainfort, P. Moururat, P. Pepin, J. Petit, G. Cabane and M. Salasse, M~m. scient. Revue Mbtall. 60, 125 (1963). 13. P. Morgund, P. Moururat and G. Sainfort, Acta metall. 16, 867 (1968). 14. K. M. Chang, Metall. Trans. 21A, 3027 (1990). 15. H. Xiao and I. Baker, Scripta metall, mater. 28, 1411 (1993).

Intermetallic Alloys H, Proc. o f the Material Research Society, Pittsburgh, Vol. 81, p. 315 (1987). 11. M. A. Crimp and K. M. Vedula, Mater. Sci. Engng 78,

1730

BAKER et al.: THE FLOW AND FRACTURE OF FeAI

16. D. J. Gaydosh, S. L. Draper, R. D. Noebe and M. V. Nathal, Mater. Sci. Engng AIS0, 7 (1992). 17. P. Nagpal and I. Baker, Mater. Char. 27, 167 (1991). 18. A. Lefort, U. Franzoni, O. Tassa, J. Le Coze, O. Cayla and A. Magnee, Proc. 6th Japanese Int. Syrup. on Intermetallic Compounds (edited by O. Izumi), Japan Institute of Metals, Sendai, p. 579 (1991). 19. I. Baker and D. J. Gaydosh, Physica status solidi (a) 96, 185 (1986). 20. P. Nagpal and I. Baker, Metall. Trans. 21A, 2281 (1990). 21. O. Klein and I. Baker, Scripta metall, mater. 27, 1823 (1992). 22. N. Junqua, J. C. Desoyer and P. Moine, Physica status solidi (a) 18, 387 (1973). 23. C.T. Liu, E. H. Lee and C. G. Mckamey, Scripta metall. 23, 875 (1989). 24. D. J. Gaydosh and M. V. Nathal, Scripta metall, mater. 24, 1281 (1990). 25. Y. A. Chang, L. M. Pike, C. T. Liu, A. R. Bilbrey and D. S. Stone, J. 1ntermet. 1, 107 (1993). 26. I. Baker and P. R. Munroe, in High Temperature Aluminides and lntermetallics (edited by S. H. Whang, C. T. Liu, D. P. Pope and J. O. Stieger), p. 133. TMS, Warrendale, Pa (1990). 27. M. Nakamura and Y. Saka, J. Mater. Sci. 23, 4041 (1988). 28. T. Takasugi and O. Izumi, J. Mater. Sci. 23, 1265 (1988). 29. G.W. Ardley and A. H. Cottrell, Proc. R. Soc. 219, 328 (1953). 30. N. Brown, Phil. Mag. 4, 693 (1959). 31. Y. Umakoshi, M. Yamaguchi, Y. Namba and M. Murakami, Acta metall. 24, 89 (1976). 32. H. Saka, Phil. Mag. A49, 327 (1984).

33. H. Saka, M. Kawase, A. Nohara and T. Imura, Phil. Mag. 51A, 3 (1984). 34. H. Saka and Y. M. Zhu, Phil. Mag. AS1, 629 (1985). 35. H. Saka, Y. M. Zhu, M. Kawase, A. Nohara and T. Imura, Phil. Mag. A51, 365 (1985). 36. Y. M. Zhu and H. Saka, Phil. Mag. A59, 661 (1989). 37. H. Saka and M. Kawase, Phil. Mag. A49, 525 (1984). 38. M. Yamaguchi and Y. Umakoshi, Prog. Mater. Sci. 34, 1 (1990). 39. N. Broom and P. Humble, Phil. Mag. 19, 639 (1969). 40. J. T. Kim and R. Gibala, High Temperature Ordered Intermetallics IV, Proc. of the Material Research Society, Pittsburgh, Pa, Vol. 213, p. 261 (1991). 41. M.G. Mendiratta and C. C. Law, J. Mater. Sci. 22, 607 (1987). 42. P. R. Munroe and I. Baker, J. Mater. Sci. 24, 4246 (1989 ). 43. A. Fourdeux and P. Lesbats, Phil. Mag. A 45, 81 (1982). 44. N. Junqua, J. C. Desoyer and J. Grilh6, Acta metall. 30, 395 (1982). 45. N. Junqua and J. Grilh6, Phil. Mag. A 45, 81 (1982). 46. O. Klein and I. Baker, Scripta metall, mater. Submitted. 47. G. W. Groves and A. Kelly, Phil. Mag. 19, 977 (1969). 48. I. Baker and J. A. Horton, Phil. Mag. A 67, 479 (1993). 49. C. T. Liu and E. P. George, Scripta metall, mater. 24, 1285 (1990). 50. I. Baker, P. Nagpal, F. Liu and O. Klein, The Processing, Properties and Applications of Metallic and Ceramic Materials (edited by M. H. Loretto and C. J. Beevers), Vol. II, p. 665. MCE Publications Ltd, U.K. (1992). 51. M. M. Shea and N. S. Stoloff, Metall. Trans. 5, 755 (1974). 52. T. L. Johnston, R. G. Davies and S. Stoloff, Phil. Mag. 12, 305 (1965). 53. P. R. Munroe and I. Baker, Acta metall, mater. 39, I01 i (1991).