AB

Synthesis of Iron Aluminide-Al,O, Composites by In-Situ Displacement Reactions

c o d F - - P p $ ~ ~ - - 4 R. Subramanian, C.G. McKamey, L.R. Buck and J.H. Schneibel

Metals and Ceramics Division Oak Ridge National Laboratory

TRACT

Composites consisting of an iron aluminide matrix with ceramic particle reinforcements, such as alumina,

could improve the high temperature strength without compromising the oxidation resistance. In this paper, the

feasibility of processing Fe-A1 alloy/Al,O, composites by an in-situ displacement reaction between Fe-40 at.% Al

and iron oxide, Fe,O,, is investigated. Simple powder metallurgical processing was performed without resorting

to any externally applied pressures or deformations during the high temperature processing step. The

microstructural features of the composites are rationalized based on results from diffusion couples. Preliminary

mechanical properties such as fracture toughness, yield strength and hardness are determined and compared with

the values obtained for monolithic iron aluminide - Fe-28 at. % Al. Results suggest that a significant improvement

in the properties is needed and further avenues for modifications, such as changes in the interface strength and

externally applied forces during processing, are suggested.

INTRODUCTION

Iron aluminide intermetallic alloys are of great interest in applications requiring high resistance to oxidation and

sulfidation [ 1,2] and are unequaled in certain molten salt applications [3]. They are also less expensive than many

currently-used high-temperature alloys, as they contain no nickel and only minor amounts of other alloying

elements 111. Recent development efforts have resulted in alloys with improved room temperature properties [ 11.

However, a decrease in strength at temperatures above 873 K have restricted the use of these alloys to lower

temperatures. The advantages they offer in terms of corrosion resistance have therefore not been fully exploited.

'The submitted manuscript has been authored by a contractor of the US. Government under contract No DE-ACOS 460R22464. Accordmgly. the US. Government retains a nonexclusive. royalty-free lwnse to publsh or reproduce the pub)ished tom of thls contnbulm, or allow others lo do so. tor

~~~~%~~~

1 U.S. Government purposes.

It is well-known that reinforcement of metals, including intermetallic alloys [5,6], with ceramic phases to form -

a metal-matrix composite can improve their high temperature strength. To date, only a few composites of FeAl or

Fe3A1 with ceramic reinforcing phases have been fabricated and studied. The fabrication routes have included arc-

melting of monolithic FeAl with Ti and B to produce a dispersed TiB, phase [7], hot isostatic pressing of

mechanically-alloyed powder mixtures of FeAl with Y203 [SI, and processing of oxide-dispersion-strengthened

(ODS) alloys of FeAl with Y20, [9]. When the reinforcement phase is mechanically mixed into the matrix phase,

. it may not be in thermodynamic equilibrium with the matrix at high -temperatures and can thus ‘degrade the

properties. In-situ produced reinforcement phases are more likely to be thermodynamically compatible with the

matrix. Processing of composites by displacement reactions allows the growth of reinforcements in-situ, which is

becoming an increasingly important concept in composite synthesis [ 1 1 - 15 1. By appropriately selecting the reactants and minimizing the volume changes resulting due to the reaction, near net-shape metal and intermetallic-

oxide composites can be obtained.

Displacement reactions are phase transformations between two or more elements or compounds resulting in

the formation of new product compounds that are thermodynamically more stable than the starting reactants [ 1 11.

Although these reactions occur in the solid state, they are similar to the Thennit reaction [16] in that they are

exothemric and their progress to completion depends upon the thermodynamics and kinetics of the starting and

ending phases. The product phases exhibit specific morphologies that depend on the relative stabilities of the

growth interfaces of the product phases [ 1 11. While some systems demonstrate stable planar or layered growth

features throughout the reaction. other systems exhibit morphological instabilities that cause the product phases to

interpenetrate the parent phase during growth from initially planar interfaces. Therefore, displacement reactions

exhibiting morphological instabilities can be used to grow interpenetrating reinforcement phases in situ.

Recently. a number of composites for structural applications have been processed by using displacement

reactions [ 12- 151. Henager et. ai. have demonstrated the feasibility of producing Sic-reinforced MoSi, and Ni-Al

alloy/Al,O; composites by solid state displacement reactions [ 121. Wi-AI alloy/ A1,0, composites with an

interpenetrating microstructure have also been processed by a reaction between NiO and AI [ 12,131. Processing of

NiAI(+ dissolved Ti) or NiAl (+ dissolved Si) contgning Al,O, precipitates by displacement reactions between

NiAl and TiO, or NiAl and SiO, has been explored [14]. Also, displacement reactions between liquid Al and

2

6 mullite has resulted in a net shape processing of a metal-ceramic composite with an interpenetrating microstructure

of A1203 and A1 [ 151.

In this paper, preliminary results from the processing of iron aluminide (Fe-AI alloys) reinforced by alumina

particles are discussed. The main objective of this research is to evaluate the feasibility of using displacement

reactions to process these composites. Details of the processing steps are discussed and microstructural

characterization is performed. The morphology of the microstructure is consistent with results from interdiffusion

studies of the reactants. Preliminary results from the room temperature evaluation of fracture toughness, flexural

strength and hardness are presented.

EXPERIMENTAL DETAILS

The starting materials for the processing of the composites were Fe-40 at. % Al powder (-325 mesh,

Homogeneous Metals Inc.) and Fe,03 powder (Johnson Matthey, 1-5 pm). Iron aluminide (Fe-40 at. %AI) and

Fe,O, powders were mixed in two separate molar ratios - 9:l and 8:3. The powders were dry mixed in a NalgeneTh' bottle for 8 hours and then isostatically pressed at a pressure of 350 MPa into compacts of about 5 cm

in diameter and 1 cm in height. The pressed pellets were placed on an alumina dish and sintered in a furnace for 1

and 3 hours, at temperatures ranging from 1100 - 1400 "C. The final successful processing technique required a

two-step heating cycle. After an initial sintering for 60 mins at 1450 "C the temperature was increased above the

melting point of the product iron aluminide matrix allowing densification due to the liquid phase. The mixture with

the reactant ratio of 9: 1 was heated to 1450 "C and that with the reactant ratio of 8:3 was heated to 1500 "C. The

temperature was maintained for only 15 mins above the melting point. The samples were allowed to cool to room

temperature and then were cut for metallographical examination and evaluation of mechanical properties. The

densities of the sintered samples were determined by Archmedes principle. Microstructures and fracture surfaces ,

were observed using a Hitachi S4100 field emission gun scanning electron microscope. An m>S detector was

utilized to determine the composition of the matrix. In order to better understand the microstructurd features in the

composite. diffusion couples between iron oxide Fe20, and iron aluminide were prepared and reacted at 1500 "C

for 3 hours. A sintered piece of Fe,O, and an as-cast piece of Fe40A1, cut into appropriate sizes, were used as

reactants in the diffusion couple. The reaction products were analyzed using an EDS detector.

For a preliminary evaluation of the mechanical properties, room temperature hardness, fracture toughness and

yield strengths were evaluated. Hardness of the composites was determined using a Buehler Micromet 2001

microhardness tester with a Vicker's indentor and a load of 500 g. Fracture toughness was measured by a

chevron notch technique using a 3-point bend test set-up. Using the same test set-up a load displacement curve

was obtained for the composites and as-cast iron aluminide (Fe-28 at. % Al) and a yield strength was determined

at a value of 0.2 % strain.

RESULTS AND DISCUSSION

Before discussing the results, it is important to review the information regarding phase stability limits of

various oxides in the ternary system Fe-A1-0. Fig. 1 is a schematic representation of an isothermal section in the

Fe-Al-0 system at about 1200 "C and it shows the presence of binary oxides, namely, FeO, Fe,O,, Fe,O, and

AI-0, - -. and one ternary oxide FeAl,O, [ 171. Another ternary compound FeAlO, also exists, but it is found to be

stable only at temperatures between 1318 "C and 1500 "C and at oxygen partial pressures greater than 0.5 atm

[ 181. The locations of the final compositions in the ternary system after reaction between Fe-40 at.% A1 and Fe,O,

powders for two different ratios of the reactants are shown in Fig. 1. ID1 and ID2 denote composites obtained by

reaction of Fe4OA1 and Fe-0, - . in molar ratios of 9: 1 and 8:3, respectively. The two reactions are shown below in

Eqs. ( 1 ) and ( 2 ) . The alumina contents in the composites produced by reactions (1) and (2) are calculated to be

7.25 vol. 5% and 20.8 vol. %, respectively.

9 Fe,AI, . - + Fe,O, + 45Feo,,,A10,3,, + A1,0,

8 Fe,AI, . - + 3Fe-0, - . + 40Feo~,,Alo,,, + 3A120,

4

- Experimental Results Reactions (1) and (2) were carried out at temperatures ranging from 1000 to 1400 "C for 1 or 3 hours under a

dynamic vacuum of about 10" Pa to identify the optimum processing condition for obtaining dense samples with

homogeneous microstructures. From SEM observations shown in Fig. 2a and 2b, it is evident that Al,O1 - - (the

darker phase) was present around the iron aluminide grains (white regions) for both ID1 and ID2 compositions.

Typical densities that were about 85 96 of the theoretical density for ID1 and 87 9% for ID2. Further improvements

in density can be obtained by utilization of external pressures such as hot pressing or forging. But, in order to

keep the processing technique simple and inexpensive, a liquid phase sintering step was introduced after the

reaction between iron aluminide and Fe,O,. In a single heat treating cycle, after reaction between the powders at

1300 "C for 1 hour, the temperature was increased to 1450 "C for the sample denoted as ID1 and to 1500 "C when

sintering the samples denoted as ID2. From reactions (1) and (2) it is clear that the intermetallic grains in ID2 have

a higher iron concentration than ID1. Since an increase in iron content increases the melting point, higher liquid

phase sintering temperatures are needed. The specimens were maintained at these temperatures for 15 mins. The

densities of the composites increase from 85 % to 96 96 of the theoretical density for ID1 and, from 87 ?6 to 95 %

of the theoretical density for ID2. Figures 2c and 2d show the changes in the microstructures after liquid phase

sintering. For IDI, due to a lower volume fraction of Al,O, precipitates, upon melting of the iron aluminide

grains, the precipitates get more dispersed within the composite. For ID2, the change in the microstructural

features is not very significant.

. The alumina particles are formed as rings around the iron aluminide grains, probably from a reaction of

Fe4OAI with Fe,O, present along the grain boundaries. X-ray diffraction results show the presence of Al,O, - - and

Fe-rich AI alloys. No peaks corresponding to FeAl,O, were detected. This is in agreement with results from

investigations by Wolski et. al. [19], where for sintering temperatures greater than 1200 "C, FeAl,O, was

identified to be unstable in contact with FeAl. EDS analysis shows that two different compositions are obtained

for the iron aluminide matrix depending on the two ratios of the reactants. For the mixture with the Fe40Al and

Fe-0, ratio of 9:1, the matrix composition was Fe0,,,A1,,, and for that with the ratio of 8:3, the matrix

composition was Fe,,,Al0,,. These values indicate that the aluminum concentrations are lower than expected from

the reactions ( 1 ) and (2), possibly due to evaporation of aluminum at the high processing temperatures. To better

5

understand the reactions between the powders, diffusion couples between the reactant phases were set up. Also, ./ based on the morphology of the reaction interface in a diffusion couple, the microstructural features of the

~ composite can be rationalized.

A cast piece of Fe4OAI and a sintered pellet of Fe,O, were reacted at 1350 "C in a dynamic vacuum of about

lo-' Pa. Fig. 3 shows the reaction layer obtained in the diffusion couple. Typically, after reaction the slabs of the

reactants never held together. They were glued together later and then mounted for typical metallographic

preparation. This region shows up as a large crack in the figure. Al,O, phase is formed as a reaction product and

the reactant slab Fe,O, is transformed to the spinel phase Fe,O,.

A qualitative explanation for the reaction products can be presented based on the information available

regarding the stability limits of the oxides under varying oxygen partial pressures (see Fig. 4 [ 181). It is assumed

that the oxygen partial pressures are those established by a local equilibrium at the reaction interfaces. Line AB

represents the diffusion couple between Fe40Al/Fe,03. From simple mass balance considerations during diffusion

. processes, the reaction path should lie on either side of the line AB. For line AB, the reaction path could follow

the sequence Fe,O,, FeAlO,, A1,0,, Fe,A1 and Fe40A1. Ths is in ]?artial agreement with reaction products

observed for the Fe4OA1/Fe,O3 diffusion couple. The concentrations of Fe and AI in the intermetallic phase do not

have a value equal to that of Fe,A1 and no concentration gradients can be detected. This is due to easy

homogenization of the Fe and Al concentrations in the intermetallic slab at 1350 "C. The final composition of he

intermetallic phase in the diffusion couple was Feo,,,Al0 ,,. In the Fe40Al/Fe20, diffusion couple, two additional discrepancies from the expected reaction products are observed - no FeAIO, is detected and the Fe,O, reactant is

transformed to Fe,O,. FeA10, is stable above 13 18 "C and also when oxygen pressures are greater than 0.5 atm.

However, the diffusion couple is reacted in a dynamic vacuum and FeA10, may not be stable at the lower oxygen

pressure. The use of a dynamic vacuum is very likely the reason for the reduction of Fe,O, to Fe,O,. On the other

hand. if the formation of Fe,O, were a result of the reduction reaction by iron aluminide, then for further

formation of A1,0, the diffusion path should traverse through the FeA120, spinel field and this phase is not

observed from our experiments.

6

ffusivities of vill ., The morphology of the reaction interface is determined by differences in t,e lOUS

components and also their activity gradients through the reaction products, as explained in Ref. [ 191. Based on the

results of displacement reactions in oxide and sulfide systems, Van Loo et al. reported that a layered morphology

in a reaction couple can be obtained under two circumstances [ 191. Following a similar analysis, a planar interface

can be obtained either by formation of Fe, or M,O, next to the iron oxide slab. Formation of Fe must proceed

through a series Fe,O,, Fe,O,, FeO and then Fe [203. However, FeO and Fe phases are not- observed in our

study. Formation of A1,0, next to the oxide reactant is possible if diffusion of oxygen through A1,0, i's very slow

compared to that of the exchange if Fe and A1 atoms. This is a reasonable assumption, since the oxygen atoms in

many oxides form a rigid oxide sublattice with predominantly cationic diffusion. In such a case, a layered

morphology, with Al,O, formation next to the iron oxide, can be rationalized. From this analysis, an

interpenetrating microstructure is expected from reacting a mixture of Fe40AI and Fe,O, powders and this agrees

well with the experimental observations (see Fig. 2d). .

Mechanical Properties

Evaluation of the mechanical properties included the determination of hardness, fracture toughness and yield

strength by 3-point bend testing at room temperature and the results are reported in Table 1. The processing of

these composites involved a furnace cool from either 1450 "C or 1500 "C and a large concentration of vacancies

could be retained, increasing the hardness in monolithic FeAl-based iron aluminide [21]. It is difficult to correlate

the trend in hardness with the properties of the constituents due to a simultaneous change in the matrix

composition and alumina volume fraction. Hardness measurements after different heat treatments, in order to

anneal the vacancies out, are underway. Observations of the indentations in an SEM revealed that there were no

cracks visible in the matrix of either composite (see Fig. 5). Results from 3-point bend testing provided the load-

displacement data shown in Fig. 6. Using an analytical expression relating the crosshead displacement to the strain

in a bend bar [22], values for the yield strength at 0.2 96 offset strain can be determined. These values are shown,

along with the fracture stress data in Table 1. It is clear that with the presence of alumina particles the yield

strength decreases significantly. This could be due to a poor interface strength between alumina and iron aluminide

which results in easy debonding. Also, any pores present along the interface could act as crack initiating flaws.

7

Fracture toughness was determined by using chevron-notched specimens in a 3-point test set-up. The toughness

of the monolithic iron alurninide was significantly higher than those of the composites. Fracture toughness of the

iron aluminide matrix with 7 vol. % Ai,O, particles (IDl) was about 32 MPa.m'" whereas with the increase in the

alumina content to 20 vol. %, the fracture toughness decreases to 8-10 MPa.m"2. The fracture surfaces shown in

Fig. 7 clearly indicate significant debonding at the aluminahon aluminide intermetallic interfaces. Imprints of the

alumina particles on the iron aluminide particles, possibly formed during the liquid phase sintering phase of the

processing technique are clearly visible on the fracture surfaces, indicating the lack of good wetting. Bridging of

the cracks by deformation of the iron aluminide phase was not observed and this could be one of the reasons for

the low fracture toughness for the composite with 20 vol. % alumina.

Clearly, significant improvements in the mechanical properties are required for structural applications. It is

clear, from the observations of fracture surfaces, that there is an extensive debonding between the alumina

particles and iron aluminide matrix. Improvements in the interface strength may be possible by additions of small

amounts of other elements, such as chromium, which is also known to increase the ductility of iron aluminide

[23]. An improvement in the interface strength and ductility of the iron aluminide matrix could aid in the

improvement of both fracture toughness and yield strength. Further improvement in processing by hot pressing or

HIPing can also lead to higher densities and better properties.

SUMMARY

In this paper. it was shown that iron aluminide matrix reinforced with alumina particles can be processed by an

in-situ displacement reaction between Fe40Al and Fe,O,. Depending upon the ratios of the two reactants, two

composites - one with Fe- 32 at. 96 A1 matrix with 7 vol. 9h alumina and another with Fe- 20 at. 96 Al and 20 vol. '3.1 alumina - can be processed. The composites are processed by a simple technique to yield an almost fully dense

material by heat treating in two stages. The first stage involves the formation of the composite by reaction at

1300 "C for 1 hour and then heating it above the melting point of the product iron aluminide to liquid phase sinter

the product. The composite with 20 vol. % had a yield strength of 73 MPa, hardness of 222 H,, fracture

toughness of 8 MPa.m". Improvements in the bend strength and fracture toughness is expected with changes in

8

the interface strength by alloying additions such as chromium. Hot pressing or forging of the composite would be

essential for obtaining samples with higher density.

ACKNOWLEDGMENTS

The authors thank Dr. Phil Maziasz and Dr. Bruce Pint for a thorough review of the manuscript. This research

is sponsored by the Laboratory Directed Research and Development Program of the Oak Ridge National

Laboratory, and by the Division of Materials Sciences, U.S. Department of Energy, under Contract No. DE-

ACO5-96OR22464 with Lockheed Martin Energy Research Corp. This research was also supported in part by an

appointment to the ORNL Post-Doctoral Research Associates Program administered jointly by the ORISE and

ORNL. L.R. Buck was supported through the American Chemical Society’s Project SEED (Summer Education

Experience for the Disadvantaged) at the Oak Ridge National Laboratory.

DISCLAIMER

This report was prepared as an a m u n t of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsi- bility for the accuracj, completeness, or usefulness of any information, apparatus, product, or proccss disclosed, or represents that its use would not infringe privately owned rights. Refer- ence herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, rccom- mendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

10

REFERENCES

1. C. G. McKamey, J. H. DeVan, P. F. Tortorelli, and V. K. Sikka, J, Mater. Res. 6, 1779 (1991).

2. J. H. DeVan and P. F. Tortorelli, Corrosion Science 35(5-8), 1065-71 (1993).

3. P. F. Tortorelli and P. S. Bishop, p. 91 in Environmental Eflects on Advanced Materials, R. H. Jones and

R. E. Rickers, eds., The Minerals, Metals and Materials Society, Warrendale, PA, 1991.

4. C. T. Liu, E.H. Lee and C.G. McKamey, Scripta Met. 23,875 1989.

5. C . G. McKamey, J. A. Horton, and C. T. Liu, p. 32 1 in High-Temperature Ordered Intermetallic Alloys 11,

M R S Proc. Vol. 81, eds. N. S . Stoloff, C. C. Koch, C. T. Liu, and 0. Izumi, MRS, Pittsburgh, PA, 1987.

6. J. H. Schneibel, P. Grahle, and J. Rosler, Mat. Sci. & Eng. A153, 684 (1992).

7. B. H. Rabin, R. N. Wright, and J. R. Knibloe, Mat. sci. & Eng. A153,706 (1992).

8. I. G. Wright, C. G. McKamey, and B. A. Pint, p. 355 in Proc. qhAnnual Conference on Fossil Energy

Materials, ed. N . C. Cole and R. R. Judkins, ORNL, ORNL/FMP-95/1 (August 1995).

9. A. U. Seybolt, Trans. ASM 59, 861 (1966).

10. R. A. Rapp, A. Ezis, and G. J. Yurek, Met. Trans. 4, 1283 (1973).

1 1. C. H. Henager, Jr., J. L. Brimhall, and L. N. Brush, Mat. Sci. & Eng. A195, 65 (1995).

12. S.A. Jones. J. M. Burlitch, J. Mat. Sci. Letters, 41(2), 1994.

13. P. Krishnan, A. Costa e Silva, and M. J. Kaufman, Scripta Metall. Mater. 32,839 (1995).

14. M. C. Breslin, J. Ringnalda, L. Xu, M. Fuller, J. Seeger, G. S. Daelhn, T. Otani, and H. L. Fraser, Mat.

Sci. & Eng. A195, 113 (1995).

15. F.A. Elrefaie and W.W. Smeltzer, Met. Trans. 14B, 85-93 (1983).

16. R. L. O’Brien, ed., Welding Handbook, pp. 892-900 in Vol. 2 Welding Processes, American Welding

Society, Miami, FL, 1991.

1 7. Ternan, Alloys, A Comprehensive Compendium of Evaluated Constitutional Data and Phase Diagrams,

Eds. G. Petzow and G. Effenberg, Vol. 5, pg 346, 1971.

18. K. Wolski, F. Thevenot and J. LE Coze, Mat. Sci. Eng. A207, 105-1 14 (1996).

19. F.J.J. van Loo, J. A. van Beek, G. F. Bastin, and R. Metselaar, Oxid. Metals 22(3/4), 161-80 (1984).

20. R.H. Spitzer, F.S. Manning and W.O. Philbrook, Trans. AIME, 236, 726-742 (1966).

10

21. P. Nagpal and I. Baker, Met. Trans. 21A, 2281-82 (1990).

22. W.G. Griffel, Handbook of FormuZas for Sress and Strain, pg. 16, Frederick Unger Publishing Co., 1960.

23. C.G. McKarney, J.A. Horton and C.T. Liu, Scripta Metall. 22, 1679-1681 (1988).

11

Table 1: Vicker’s hardness, yield stress and fracture stress of Fe-AI alloy/Al,O, composites

compared with monolithic Fe-28Al

Nominal Matrix Al,O, Hardness Yield Stress Fracture

Composition content (H,) (MPa) Stress (MPa)

Fe-28 at. % Al (as-cast) 0 308 463 896

Fe-32 at. % Al (ID1) 7 260 370 465

Fe-20 at. % Al (D2) 20 222 73 107

12

. .

LIST OF FIGURES

1. A schematic of an isothermal section of the ternary system Fe-AI-0 at about 1200"C, indicating the presence of

binary and ternary oxides.

2. SEM micrographs of composites processed by in-situ displacement reactions: (a) ID1 sintered for 1 hr at

1200"C, (b) ID2 sintered for 1 hr at 12OO0C, (c) ID1 sintered for 1 hr at 1300°C and then liquid phase sintered for

15 mins at 145OoC, and (d) ) ID2 sintered for 1 hr at 1300°C and then liquid phase sintered for 15-mins at 1500°C.

3. A diffusion couple of Fe-40AI and Fe,O, reacted at 1350°C in a dynamic vacuum of

4. A log P(0,) versus composition phase diagram for the Fe-AI-0 system indicating the phase stability limits of

various oxides at a temperature of 1380°C.

5. Vicker's indentation on specimen ID1 (with 7 vol.% Al,O,).

6. Comparison of bend strengths of composites, processed by in-situ displacement reactions, with monolithic

iron aluminide (Fe-28 at.% AI).

7 . Fracture surface observations of (a) ID1 - Fe,,,Fe,,, matrix with 7 vol.% AI,O, and (b) ID2 - Fe,,A10.2

matrix with 20 vol.% A1,0,.

Pa.

0

FeO

Fe A1

Fig. 1 : A schematic of an isothermal section of the ternary system Fe-A1-0 at about 1200 "C , indicating the presence of binary and ternary oxides.

Fig. 2: SEM micrographs of compositcs proccsscd by in-si tu displacement reactions (a ) ID1 sintcrcd for I hr at 12ot) 'C. (b) ID2 sintered for 1 hr at 12OU "C, (c) ID1 sintcrcd for I hr at 1300 'C and then liquid phasc sintcrcd for 15 mins at1450 'C and (d) ID2 sintcrcd for 1 hr at 1300 'C and then liquid phasc sinteredfor 15 mins at 1500 "C

A

Iron A I u minide Fe0.654.35

50 pm

Crack -

/ *l2*3

Fig. 3: A diffusion coupic of Fe4OAl and Fc703 reacted at 1350 "C in a dynamic vacuum of 1 0 - 3 pa.

FeAlO, / Fe701

\

\ . \ - 1

1 I i I I

C 3.2 0 4 0.6 o.a 1 .o

Fig. 4: A log P(02) versus composition phase diagram for the Fe-AI-0 system indicating the phase stability limits of various oxides at a temperature of 1380 "C.

. Fig. 5 : Vicker's indentation on specimen ID 1 (with 7 vol. R A1703)

t

1 . 6

1.4

1.2

1

0 . 8

0 . 6

0.4

0 . 2

0

1 1 . 1 I I I I I I I I - I 1 8 I l l I l l . Fe-28 at. % A1 -

- Room temperature 3 - point bend test results (20 mm span, -

- crosshead speed =

- - - -

- - - - - - - - .

0 0 . 2 0 . 4 0 . 6 0 . 8 1 1 . 2 Displacement (mm)

Fig. 6: Comparison of bend strengths of composites, processed by in-situ displacement reactions, with monolithic iron aluminide (Fe- 28 at. % Al)

Fig. 7 : Fracture surf'acc observations 0 1 (a) ID1 - Feo.,XA10,~7 matrix with 7 vol. '3 A1203 and (h) ID2 - Fc,,,HAI,,.2 matrix with 20 vol. c7( Al7O7 - .