State of the Art: In the discussion of the state of the art that follows, reference is made to certain structures and/or methods. However, the following references should not be construed as an admission that these structures and/or methods constitute prior art. Applicant expressly reserves the right to demonstrate that such structures and/or methods do not qualify as prior art against the present invention.

Ordered intermetallic alloys based on iron aluminide Fe3Al offer a combination of attractive properties such as excellent resistance to oxidation and sulfidation at high temperature, and high strength to weight ratio. (See, for example, R. S. Sundar, et al., Mater. Sci. and Eng. A, 258, (1998) 219-228.) They are also potential low-cost replacements for more expensive high temperature structural alloys containing strategic elements such as nickel and chromium. Although iron aluminide alloys exhibit limited room temperature ductility, poor high temperature strength, low fracture toughness, poor machinability and poor resistance to environmental embrittlement, substantial improvement in these properties can be achieved by a combination of composition and process control.

More recently, intermetallics based on nickel, iron and titanium aluminides have been the subject of research due to their excellent thermal stability at high temperatures coupled with their unique combination of properties such as low densities and good room temperature and high-temperature tensile strengths. (See M. R. Hajaligol et al., Mater. Sci. and Eng. A, 258 (1998) 249-257.) Of the intermetallics, iron aluminides based on FeAl with a B2 structure (ordered BCC structure with aluminum atoms occupying the body centers) are of more interest than Fe3Al-based alloys. Alloying these alloys with Mo, Zr and others results in a combination of attractive properties such as oxidation, corrosion and sulfidation resistance at high temperatures. Additionally, the alloy possesses reasonable strength at high temperatures for use as a structural material. The room temperature ductility of FeAl alloys are generally in the range of 2-6 %, and the elongations are influenced by room temperature embrittlement.

The low ductility of FeAl alloys necessitates hot working of cast materials at high temperatures, and hot working approaches limit the manufacturability of sheets and rods.

Metallurgical processing techniques based on melting and casting, forging and rolling have been used to understand the processability of FeAl alloys. Casting techniques such as Air Induction Melting (AIM) and Vacuum Induction melting (VIM) have generally proven unsatisfactory due to the high porosity of ingots processed by these techniques. However, AIM and VIM ingots subjected to a secondary Vacuum Arc Melting (VAM) process have exhibited improved porosity and reduced defects. (See C. Testani, et al., Proc. Inter.

Symposium on Nickel and Iron Aluminides: Processing, Properties, and Applications, Eds. S.

C. Deevi, P. J. Maziasz, V. K. Sikka, and R. W. Cahn, ASM International, Materials Park, OH, 1996, pp. 213.) Ingots subjected to hot working display properties that are temperature and strain rate dependent, thereby necessitating processing at strain rates lower than commercially viable to avoid unwanted properties such as work hardening. This may be partially overcome by extrusion in which low strain rates, in combination with preheated dies and hydrostatic pressure, combine to avoid brittle fracture and yet control detrimental grain growth.

However, all of the above processes suffer from complexity and cost.

It has been shown that the yield strengths, ultimate tensile strengths and tensile elongations of hot extruded rods of FeAl alloys are superior to the properties of the cast FeAl alloys due to the fine grain microstructure of the hot extruded FeAl alloys. (See P. J.

Mazsiaz, et al., Proc. Inter. Symposium on Nickel and Iron Aluminides: Processing, Properties, and Applications, Eds. S. C. Deevi, P. J. Maziasz, V. K. Sikka, and R. W. Cahn, ASM International, Materials Park, OH, 1996, pp. 157.) In addition, it has been shown that reactive hot extrusion of Fe and Al powders improves the properties of fine grained Fe-24 wt. % Al over the properties of cast materials. (See S. C. Deevi, et al., Proc. Inter.

Symposium on Nickel and Iron Aluminides: Processing, Properties, and Applications, Eds. S.

C. Deevi, P. J. Maziasz, V. K. Sikka, and R. W. Cahn, ASM International, Materials Park, OH, 1996, pp. 283.) For commercial viability, powder metallurgy is promising since finer microstructures with a uniform distribution of micro-constituents can theoretically be produced. These, by analogy to other fine grained FeAl alloys, should also exhibit improved properties. However, the critical processing paths to achieving these alloys has not been clearly delineated.

In view of the above, it is desirable to obtain a simple and cost effective processing route to produce iron aluminides with fine grain microstructures and which ensures high quality defect-free products on a repeatable basis.

Due to the irregular shapes of water-atomized powders, they can be cold compacted and continuously rolled to produce good quality sheet products. In contrast, gas atomized powders have spherical shapes and require hot compacting methods to consolidate them. Hot extrusion of these powders has resulted in semi-products with fine-grained microstructures with attractive properties. For developing components from these semi-products, it is important to characterize their hot deformation behavior so as to help in designing metal working processes.

Different approaches available for evaluating the hot working mechanisms include evaluation of the shapes of the stress-strain curves, the standard kinetic parameters and the processing maps. The kinetic analysis of hot deformation uses the rate equation: é =Xoçexp [-Q/RT] (1) where is the strain rate, A is a constant, o is é the flow stress, 71 is the stress exponent, Q is the activation energy, R is the gas constant and T is the temperature. The kinetic parameters-the stress exponent and the apparent activation energy-are evaluated from experimental data and compared with the corresponding values known for some atomistic mechanisms for identifying the rate controlling one. The limitations of this analysis have been its applicability to complex alloys and its inability to optimize the hot workability.

The use of processing maps for optimizing hot workability and controlling microstructure during mechanical processing has been found to be very beneficial. (See, for example, Y. V. R. K. Prasad and S. Sasidhara, Eds., Hot Working Guide: A Compendium of Processing Maps, ASM International, Materials Park, OH, 1996.) With the help of a processing map, it is possible not only to arrive at the optimum parameters for designing a metalworking process without resorting to expensive and time consuming trial and error methods but also to control the grain size.

In brief, depicted in a frame of temperature and strain rate, power dissipation maps represent the pattern in which the power is dissipated by the material through microstructural changes. The rate of this change is given by a dimensionless parameter called the efficiency, il, of power dissipation: il= 2m/ (m + 1) (2) where m is the strain rate sensitivity of flow stress. Over this frame is superimposed a continuum instability criterion for identifying the regimes of flow instabilities, developed on the basis of extremum principles of irreversible thermodynamics as applied to large plastic flow and given by another dimensionless parameter: [aln (mlm + I) Ialne] +in (3) wherein is the applied strain rate. Flow instabilities are predicted to occur when is negative. These two maps together constitute a processing map, which exhibits domains with local efficiency maxima representing certain specific microstructural mechanisms and also regimes of flow instabilities.

The exploitation of powder metallurgy routes is very promising since a much finer microstructure with a uniform distribution of microconstituents can be produced by this technique. In addition, the mechanical processing of compacts formed by powder metallurgical methods is an important first step in view of the limited workability of the powder metallurgy compacts.

BRIEF SUMMARY OF THE INVENTION The present invention overcomes the deficiencies previously associated with conventional casting and powder metallurgical technologies for the processing of intermetallic alloys.

In addition, the invention provides a method of manufacturing a worked product from an intermetallic alloy such as iron, nickel or titanium aluminide alloy which produces sound material retaining all of the advantageous material properties commonly associated with these alloys while providing the added advantage of lower cost.

More particularly, the present invention provides for a method of manufacture comprised of preparing a body from an intermetallic alloy powder, preferably an iron aluminide, the powder formed by either water atomization or gas atomization techniques, and hot working the body at a strain rate of 0.001 to 1.0 s-1 and at a temperature range above 750°C during which the intermetallic alloy undergoes grain refinement. Alternatively, the body is forged or rolled at a temperature of 1100 to 1250°C prior to the hot working step.

The strain rate, the selection of which is aided by the use of power dissipation maps, is sufficient to achieve either dynamic recrystallization or superplastic deformation of the intermetallic alloy.

BRIEF DESCRIPTION OF THE DRAWING FIGURES The objects and advantages of the invention will become apparent from the following detailed description of preferred embodiments thereof in connection with the accompanying drawings in which like numerals designate like elements and in which: Figure 1 is a flow diagram illustrating the steps of the method to manufacture a worked product from water atomized powders of an intermetallic alloy.

Figure 2 is a flow diagram illustrating the steps of the method to manufacture a worked product from gas atomized powders of an intermetallic alloy.

Figures 3 a-b show surface morphologies of (a) water atomized (WA), and (b) gas atomized (GA) powders.

Figure 4 shows an initial microstructure of a compacted and extruded billet of water atomized Fe-24 weight % A1 alloy.

Figure 5 shows an initial microstructure of a compacted and extruded billet of gas atomized Fe-24 weight % Al alloy.

Figure 6 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 600°C obtained in compression at different strain rates.

Figure 7 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 650°C obtained in compression at different strain rates.

Figure 8 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 700°C obtained in compression at different strain rates.

Figure 9 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 750°C obtained in compression at different strain rates.

Figure 10 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 800°C obtained in compression at different strain rates.

Figure 11 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 850°C obtained in compression at different strain rates.

Figure 12 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 900°C obtained in compression at different strain rates.

Figure 13 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 950°C obtained in compression at different strain rates.

Figure 14 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 1000°C obtained in compression at different strain rates.

Figure 15 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 1050°C obtained in compression at different strain rates.

Figure 16 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 1100°C obtained in compression at different strain rates.

Figure 17 shows true stress-true plastic strain curves of an iron aluminide alloy made from WA powder at 1150°C obtained in compression at different strain rates.

Figure 18 shows true stress-true plastic strain curves of an iron aluminide alloy made from GA powder at 850°C obtained in compression at different strain rates.

Figure 19 shows true stress-true plastic strain curves of an iron aluminide alloy made from GA powder at 900°C obtained in compression at different strain rates.

Figure 20 shows true stress-true plastic strain curves of an iron aluminide alloy made from GA powder at 950°C obtained in compression at different strain rates.

Figure 21 shows true stress-true plastic strain curves of an iron aluminide alloy made from GA powder at 1000°C obtained in compression at different strain rates.

Figure 22 shows true stress-true plastic strain curves of an iron aluminide alloy made from GA powder at 1050°C and at different strain rates.

Figure 23 shows true stress-true plastic strain curves of an iron aluminide alloy made from GA powder at 1100°C obtained in compression at different strain rates.

Figure 24 shows true stress-true plastic strain curves of an iron aluminide alloy made from GA powder at 1150°C obtained in compression at different strain rates.

Figure 25 shows the variation of flow stress with strain rate at different temperatures at a strain of 0.5 for a GA FeAl alloy.

Figure 26 is a processing map obtained on an iron aluminide alloy formed from WA powders at a strain of 0.1.

Figure 27 is a processing maps obtained on an iron aluminide alloy formed from WA powders at a strain of 0.2.

Figure 28 is a processing maps obtained on an iron aluminide alloy formed from WA powders at a strain of 0.3.

Figure 29 is a processing maps obtained on an iron aluminide alloy formed from WA powders at a strain of 0.4.

Figure 30 is a processing maps obtained on an iron aluminide alloy formed from WA powders at a strain of 0.5.

Figures 31 a-d show microstructures obtained on iron aluminide alloy specimens made from WA powders deformed at (a) 1100°C/0. 001 s-', (b) 1100°C/0. 1 sn, (c) 1150°C/0. 001 s- , and (d) 1150°C/0. 1 su.

Figure 32 shows a microstructure obtained on iron aluminide alloy specimen made from WA powders deformed at 1150°C and 100 s-1 strain rate (DRX domain).

Figures 33 a-b show microstructures of iron aluminide alloy specimens made from WA powders deformed at (a) 850°C/0. 001 s-1, and (b) 900°C/0. 1 s-1 Figure 34 relates grain size values measured on iron aluminide alloys made from WA powders deformed at different temperatures and strain rates.

Figure 35 is a schematic bifurcation diagram for iron aluminide alloy made from WA powders obtained from the changes in the processing maps with strain at a temperature of 1150°C.

Figures 36 a-c show microstructures obtained on iron aluminide specimens made from WA powders deformed at 750°C in the instability regime. (a) 100 s-1, (b) 10 s-1, and (c) 1 s-1.

Figure 37 is an Arrhenius plot for an iron aluminide alloy made from WA powders in the domain of dynamic recrystallization.

Figure 38 shows variation of average grain diameter with the Zener-Hollomon parameter for an iron aluminide alloy made from WA powders.

Figure 39 is a processing map obtained on an iron aluminide alloy formed from GA powders at a strain of 0.1.

Figure 40 is a processing map obtained on an iron aluminide alloy formed from GA powders at a strain of 0.2.

Figure 41 is a processing map obtained on an iron aluminide alloy formed from GA powders at a strain of 0.3.

Figure 42 is a processing map obtained on an iron aluminide alloy formed from GA powders at a strain of 0.4.

Figure 43 is a processing map obtained on an iron aluminide alloy formed from GA powders at a strain of 0.5.

Figures 44 a-b show microstructures obtained on iron aluminide alloy specimens made from GA powders deformed at a strain rate of 0.1 s-1 and at (a) 1050°C and (b) 1100°C.

Figure 45 shows variation of average grain diameter of an iron aluminide alloy made from GA powders with temperature at a strain rate of 0.1 s-1. The variation of efficiency of power dissipation (at g=0. 5) is also shown.

Figures 46 a-b show microstructures of iron aluminide alloy specimens made from GA powders deformed at 1100°C and at (a) 0.01 s-1 and (b) 1.0 s-l.

Figure 47 relates grain size values measured on iron aluminide alloys made from GA powders deformed at 1100°C at different strain rates.

Figure 48 shows a microstructure obtained on an iron aluminide specimen made from GA powders deformed at a strain rate of 100 s-1 and 1050°C, corresponding to the instability regime.

DESCRIPTION OF PREFERRED EMBODIMENTS Figures 1 and 2 provide schematic illustrations of the process steps employed to process intermetallic alloys from water atomized powders and gas atomized powders, respectively. Atomization techniques were used to obtain powders of FeAl alloy of either spherical or irregular shape with the target compositions (weight %) as indicated in Table 1.

Table 1 Element Al Mo Zr B C O Fe Gas atomized 24.0 0.42 0.1 0.005 0.06 0.05 Bal. Water atomized 24.0 0.42 0.1 0.005 0.06 0.31 Bal.

The surface morphologies of gas and water atomized powders are shown in Figure 3.

The water atomization technique yielded irregular shaped powders (Figure 3 A), and the gas atomization technique resulted in spherical particles (Figure 3 B).

Unlike the gas atomization technique, water atomization of FeAl alloy powders required special precautions to reduce the oxygen content of the particles and prevent formation of oxides of iron and aluminum on the surface. The oxygen content of the water atomized powder is close to 0.3 wt. % or higher. On the other hand, the oxygen content of the gas atomized powder is in the range of 0.02-0.04 wt. %, an order of magnitude lower than the water atomized powder.

The powders were filled in a steel can and extruded at 1100°C to obtain fully dense bars of the alloys. Cylindrical compression specimens of 10 mm diameter and 15 mm height were machined from the extruded material such that the compression axis is parallel to the extruded direction. The flat ends of the specimen had grooves for holding the lubricant and the edges were chamfered to avoid initial fold over.

Hot compression tests were conducted in the temperature range 600-1150°C at 50°C intervals and constant true strain rate range of 0.001-100 s-1 at intervals of an order of magnitude. The specimens were allowed to soak at the testing temperature for about 15 minutes before the start of compression. During testing, the actual temperature of the specimen as well as the adiabatic temperature rise were measured using a thermocouple inserted in a 1.0 mm hole machined at half its height to reach the center of the specimen. The specimens were coated with a borosilicate glass paste that acted not only as a lubricant but also as a protective layer to minimize oxidation. All specimens were deformed to half their height and air cooled to room temperature following deformation.

The load-stroke data obtained in compression were processed to obtain true stress-true plastic strain curves. The flow stress data obtained at different temperatures, strain rates and strains were corrected for adiabatic temperature rise, if any, by linear interpolation between log o and 1/T where o is the flow stress and T is the temperature in Kelvin. A cubic spline fit between logo and log was used to obtain the strain rate sensitivity (m) as a function of strain rate. This was repeated at different temperatures.

The efficiency of power dissipation (il) through microstructural changes was then calculated as a function of temperature and strain rate using Eq. (2) and plotted as an iso-efficiency contour map. The data were also used to evaluate the flow instability parameter () using Eq. (3) as a function of temperature and strain rate to obtain an instability map.

The starting bar and the deformed specimens were sectioned parallel to the extrusion axis and the compression axis respectively and the cut surfaces were prepared for metallographic examination.

Initial Microstructure The initial microstructure of the starting rod of water atomized material is shown in Fig. 4, which reveals an equiaxed grain structure with an average grain diameter of about 22 , um. The large grain size is expected due to the high oxygen content in this alloy. The prior particle boundaries are totally eliminated during the extrusion process due to the possible occurrence of dynamic recrystallization (DRX). The grains are slightly elongated in the extrusion direction which is horizontal in the micrograph. The microstructure also consists of a uniform distribution of finer carbide particles and larger alumina particles.

The microstructure of the starting rod of gas atomized material is shown in Fig. 5. It has an equiaxed grain structure with an average grain diameter of about 13, um. The prior particle boundaries are totally eliminated during the hot extrusion process possibly due to the occurrence of dynamic recrystallization. The microstructure also consists of fine carbide and alumina particles, which are aligned in the direction of extrusion.

Stress-Strain Behavior The true stress-true plastic strain curves obtained on samples made from water atomized powders at temperatures of 600,650 and 700°C are shown in Figures 6-8, respectively. The curves at higher strain rates (> 1.0 s-1) exhibited peaks in the flow stress at strains between 0.1-0.2. At the strain rates of 10 and 100 s-1, a second set of peaks appeared at larger strains. A visual examination of the specimens showed extensive cracking early in deformation. The occurrence of cracks is responsible for the initial drop in the flow stress, which then increases when the cracks get welded under compression. At lower strain rates, the cracking is less severe due to the compressive state of stress and the curves indicate some strain hardening. The cracking problem becomes less severe as the temperature is increased and is nearly eliminated at temperatures higher than 700°C.

The true stress-true plastic strain curves obtained on samples made from water atomized powders in the temperature range 750-1150°C are shown in Figures 9-17, respectively. These curves may be subdivided into two types depending on the strain rate. At strain rates lower than about 1. 0 s-1, the curves are essentially of steady state type although an initial drop in the flow stress occurs at temperatures higher than about 950°C. At higher strain rates, the curves exhibit a peak in the yield stress at a strain, which is higher at higher strain rates. For a strain rate of 10 s n the flow reaches a steady state at larger strains. The flow stress data obtained from the stress-strain curves at different temperatures, strain rates and strains are given in Table 2.

Table 2. Flow stress data (in MPa) on Iron Aluminide Alloy (water atomized P/M compact) obtained in compression at different temperatures, strain rates and strains (corrected for adiabatic temperature rise). Strain Strain Temperature, °C rate, s-l 750 800 850 900 950 1000 1050 1100 1150 0.001 224. 3 144.5 98. 8 66.6 53.1 37.1 26.6 12.1 13.0 0.01 330.1 245.2 168.3 114.9 87.4 65.7 43.3 31.9 23.5 0.1 388.9 337.7 208.6 185.0 150.8 110. 1 81.8 55.5 44.8 0.1 1. 0 475.3 423.0 355.7 276. 5 245.1 190.3 131. 6 102.1 69. 4 10.0 743.2 615.3 465.1 379.3 352.4 318.2 259.4 189. 4 149. 4 100. 0 840. 4 751.1 555.4 455.6 383.7 354.8 313.2 264.2 174.3 0.001 225.5 144.8 98. 5 66.9 53.8 37. 3 27.7 13. 1 12.3 0.01 331.9 246.6 169.0 115. 5 87. 4 66.3 44.5 33.1 22.7 0. 1 394.5 334.1 203. 9 182.8 147.1 107.0 80.5 55.8 43.4 0.2 1. 0 492.2 436.6 355.8 272.8 238.4 184.1 127.3 98.8 68. 8 10. 0 713. 0 596.8 447.0 362.1 326.7 282.7 236.7 154.3 127.5 100.0 853.0 815.4 613.4 496.5 411.8 376.1 330. 9 268.8 182. 5 0.001 226.8 144.5 99. 3 68.2 54.3 37.8 28.4 14.1 11.3 0.01 334. 8 248.9 172.3 117. 0 88. 2 66. 4 44.8 33.6 22. 0 0. 1 395.7 330.6 205.9 183.6 147. 1 107.6 80.6 56. 3 43. 0 0.3 1. 0 493.6 439.9 355.3 271.1 235.6 181.4 127.4 98.5 69.4 10.0 677.7 603.8 461.5 373.0 341.5 291.2 230.9 151.3 121.5 100.0 777.1 796.9 624.0 509.3 422.2 383.7 334. 1 270. 9 189.6 0.001 223.3 142.4 98.8 68. 2 53.9 38. 2 28.8 14. 5 10.4 0.01 336.9 246.6 172.3 117.5 88. 1 66.6 44.6 33.7 21.6 0. 1 395. 4 326.0 206.6 182.0 147.7 107.1 80.9 56.5 42.5 0.4 1 1 1. 0 493.3 436.7 352. 7 269.3 232.7 178.3 126.4 97.2 69.6 10.0 636. 8 577.3 457.8 373.5 336.9 290.0 223.5 151.7 121.3 100.0 683.5 753.1 605.7 501.9 421.4 382.6 327.8 261.4 189.0 0.001 216.3 139.6 97.4 66.8 54.2 37.7 28.6 14. 7 10.2 0.01 335. 8 242.9 170. 7 115. 6 86. 5 65. 1 44.0 33. 5 21.0 0.1 397.3 323.2 203.8 180.0 146.7 104.9 80. 5 56.0 41.6 0.5 1. 0 497.2 431.7 347.0 266.7 228. 7 174.3 123.7 95.8 68.8 10.0 619.9 561.5 453.9 370.9 331.0 281.4 219.4 147.8 120.6 100.0 607.7 691.6 567.1 480. 2 403.4 369.4 311.6 243.5 181. 5 The true stress-true plastic strain curves obtained on gas atomized material at temperatures between 850 and 1150°C and at different strain rates are shown in Figures 18- 24, respectively. The flow curves at strain rates lower than about 1.0 s-l are essentially of steady state type, although a stress maximum is present at lower strains. At higher strain rates, the flow curves exhibit significant flow softening after reaching a peak in the flow stress in the initial stages of deformation. Further, at temperatures higher than 1000°C and at a strain rate of 1.0 s-', the flow curves exhibit oscillations after the initial peak in the flow stress. The flow stress data obtained from the stress-strain curves at different temperatures, strain rates and strains are given in Table 3.

Table 3. Flow stress data in MPa for FeAl alloy (Gas Atomized P/M compact) obtained in compression at different temperatures, strain rates and strains (corrected for adiabatic temperature rise). Strain Strain Temperature, °C rate, s-1 850 900 950 1000 1050 1100 1150 0.001 134.0 102.9 62. 1 43.1 30.8 22. 7 16.2 0.01 188. 4 145. 2 104. 4 73.3 53.5 36.9 27.7 0.1 256. 5 215.9 153. 1 129. 7 96. 5 62. 6 43.3 0. 1 1. 0 370.6 282.7 255.3 213.8 211. 3 109.7 80. 9 10.0 551.6 444.4 427.2 338.2 284.5 210. 5 152.6 100.0 766. 5 566.8 409. 7 385. 6 330.4 260.4 210.2 0.001 132.6 97. 9 62. 9 44. 9 32. 7 23. 9 16.5 0.01 188.9 146.1 103.8 73. 0 53. 5 37. 8 28.4 0.1 253. 0 213.9 153.8 127.1 94. 5 62. 4 42.8 0.2 1. 0 366.9 276.4 247.9 204.9 204.2 106.3 79.1 10. 0 590. 8 443.4 410.6 316.5 261.0 194.4 153.4 100.0 857.7 609.0 429.6 394.7 336.4 266.4 223.0 0.001 132.8 95.6 63. 4 45. 8 33. 4 24.7 16.4 0. 01 189. 4 147.7 104.8 71.4 53.7 37.7 29. 0 0.1 252. 1 214.0 148.4 126.7 93. 9 62. 4 44.0 0.3 1.0 366.8 269.1 243.6 201.8 200.9 105.9 79.0 10. 0 592. 9 437.5 404.1 306.5 251.2 185.5 153.7 100. 0 840. 4 605.5 434.3 392.8 333.1 266.5 228.5 0. 001 128.0 90.9 62. 1 45. 5 33. 9 24. 8 16.0 0. 01 185.8 146.1 104.9 68. 1 52. 6 36. 9 28.6 0.1 245.9 213.8 142.2 126.5 92.9 62.1 44.1 0.4 1.0 359.0 265.7 237.9 197.5 197.5 103.5 77.7 10. 0 561. 0 428. 6 397.3 295.2 242.4 180.5 147.3 100. 0 781. 6 585.2 431.5 383.9 324.5 255. 7 225.9 0.001 124.1 87. 4 61.2 44.7 33.5 24.6 15.7 0.01 181.5 141.7 102.6 64. 6 51. 9 35. 7 28.5 0. 1 246. 5 209.0 135.9 123. 7 91.1 61.7 43. 8 0. 5 1. 0 351. 9 259. 1 231.8 191.7 193.1 100. 7 75.8 10.0 511.9 408. 7 384.4 277.6 230.1 169. 1 134.9 100. 0 725. 3 550.1 417. 5 364.5 308. 5 242. 7 217. 7 Comparing and contrasting the true stress-true strain curves for both water and gas atomized samples, one sees that the curves are similar. However, the gas-atomized material was harder than the water atomized, as may be expected from its finer grain size. It is not possible to arrive at the mechanism of hot deformation directly from the shapes of the stress- strain curves alone, although they may be correlated with the mechanisms if evaluated by other modeling methods. This is because several mechanisms may lead to similar shapes of stress- strain curves e. g. flow softening may suggest dynamic recrystallization (DRX), flow instability or globalarization of lamellar structures.

Water Atomized Powders (i) kinetic analysis In the hot deformation of materials, the relationship between the steady state flow stress, the temperature and strain rate is expressed by a kinetic rate equation (Eq. 1). In order to evaluate the stress exponent, the flow stress data obtained at a strain of 0.5 are plotted against the strain rate on a log-log scale at different temperatures (see Fig. 25). The value of n, which is the inverse of the slope of the line, is dependent on temperature being lower at higher temperatures. Although the data may be fitted to a straight line, it is clearly seen that the variation follows a curve, particularly at temperatures lower than 1050°C, indicating that the value of the stress exponent is strain rate dependent. Ignoring the strain rate dependence of the stress exponent, a simple activation analysis yields an apparent activation energy of about 465 kJ/mole for the gas atomized material and 430 kJ/mole for water atomized material.

These values are in agreement with those previously reported (See e. g. J. D. Whittenberger, Mater. Sci. Eng. A, 57,77 (1983) and J. D. Whittenberger, Mater. Sci. Eng. A, 77,103 (1986)), and are much higher than that for diffusion of either Fe or Al in FeAl (See e. g. M.

Eggersmann, et al., Defect and Diffusion Forum, 339,143 (1997) and E. Kentzinger, et al., Phys. Condens. Matter, 8,5535 (1996)).

(ii) processing maps The processing maps developed at strains of 0.1,0.2,0.3,0.4 and 0.5, on the basis of the flow stress data given in Table 2, are shown in Figs. 26-30, respectively. These maps are obtained by a superimposition of the instability maps over the power dissipation maps. In each of the processing maps, the regime of flow instability as predicted by the continuum criterion, given by Eq. (3), is delineated by a thick line (marked as"0") cutting across several efficiency contours belonging to the power dissipation maps. The power dissipation maps show isoefficiency contours which represent the relative rate of entropy production occurring in the material due to microstructural dissipation. They can also be termed as"microstructural trajectories"since they actually represent the rate of change of microstructure occurring during hot deformation. It is interesting to note that the curvature of the trajectories changes when the temperature is increased beyond about 950°C, the temperature at which dissolution of fine carbide particles is likely to occur. Since the material system undergoing hot deformation is dynamic, non-linear, dissipative and irreversible, it possesses the characteristics similar to those exhibiting"deterministic chaos". As the system moves towards a steady state at higher strains, the trajectories move towards microstructural attractors (domains of maximum rate of entropy production or basins of lowest dissipative energy) depending on their initial conditions. These concepts are applied to the materials system for interpreting the maps obtained at different strains.

The map obtained at a strain of 0.1 (Fig. 26) exhibits only one domain at a temperature of 1100°C and a strain rate of 0.03 s-l with a maximum efficiency of power dissipation of about 44 %. As the strain increases, this domain gives rise to another domain with a higher efficiency (i. e. 48% at a strain of 0.2 and increasing with increasing strain, see Figs. 27-30) occurring at a temperature of 1150°C and a strain rate of 0.001 s-l. This domain has a temperature range of 1000-1150°C and a strain rate range of 0.001-0.1 s-1.

Simultaneously, another domain also appears beyond a strain of 0.3 with a peak efficiency of about 38% occurring at 1150°C and a strain rate of 100 s-l. Thus, the maps given in Figs. 26- 30 suggest that the microstructure of the material is evolving during deformation as is commonly observed in systems with several state-space parameters. For example, the state- space parameters in the present case are temperature, strain rate, strain and the rate of entropy production (dissipative energy state). The best way of representing such a change is through a bifurcation diagram.

Referring to the map obtained at the strain of 0.1 (Fig. 26), the single domain observed has a high efficiency of power dissipation which suggests dissipative mechanisms like DRX.

The efficiency hill representing these contours is not a steep one since an efficiency range of only 8 % is spread over a wide temperature range (250°C). Stress-strain curves in the temperature range corresponding to this domain exhibit a small drop in yield stress suggesting that some significant softening mechanism operates initially. These features suggest that this domain represents DRX process and the changes in the maps that occur with strain give further support to this interpretation since it causes a change in the microstructure such that its response to the temperature and strain rate will change significantly. In high stacking fault energy metals, the maximum efficiency of power dissipation for DRX is about 50% while it is about 35 % in low stacking fault metals. Thus the observed value of 44 % suggests that this iron aluminide alloy has a medium stacking fault energy.

At higher strains (e. g. 0.4, Fig. 29), the processing map exhibits two domains. The domain at strain rates below about 0.1 s-1 and at temperatures above 1000°C, has a peak efficiency of about 56 % and the contours represent a steep hill (efficiency increases by about 18% within a temperature range of about 125°C). The maximum efficiency corresponds to a strain rate sensitivity of about 0.4 and the stress-strain curves are of steady-state type (Figs.

14-17). Such domains suggest the occurrence of superplastic deformation or edge cracking of the material. The results from tensile tests on similar materials have clearly shown that abnormal elongations (>300%) are obtained under these conditions, thereby confirming the occurrence of superplasticity in this domain. Typical microstructures obtained on specimens deformed at 1100 and 1150°C and strain rates of 0.001 and 0.1 s-I are shown in Fig. 31.

These exhibit very fine equiaxed grain structure. The measured average grain diameter is about 12, um, which is much finer than the initial grain size (22/mi). Also the grain size did not vary significantly with temperature or strain rate as is expected to happen in the superplasticity domain.

The higher strain rate domain (> 10 su) is probably not fully developed within the testing regime of temperature and strain rate since only a small part of it is seen in the map (Fig. 30). The stress-strain curves under conditions within the domain (Figs. 16 and 17) exhibit typical DRX features which include a peak in the flow stress followed by a steady state as well as initial oscillations reaching a steady state when the strain rate is at the lower end of the domain. Typical microstructure recorded on a specimen deformed at 1150°C and 100 s-1 (Fig. 32) exhibits wavy or irregular grain boundaries, which are considered signatures of a DRX process. Thus the high strain rate domain may be interpreted to represent a DRX process.

Representative microstructures recorded on specimens deformed at 850°C at a strain rate of 0.001 s-1, and at 900°C and 0.1 s-l are shown in Figs. 33 a-b. These temperature and strain rate combinations correspond to regions in which the microstructural changes are associated with trajectories that do not get attracted to any of the domains discussed above.

No significant change is observed in these micrographs. This is further confirmed by grain size measurements as plotted in Fig. 34. The grain refinement at temperatures higher than 1050° may be observed.

A bifurcation diagram at a temperature of 1150°C representing the changes in the deformation mechanisms occurring with strain is shown schematically in Fig. 35 which will help in understanding the changes occurring in the material with strain. Up to a strain of about 0.1, dynamic recrystallization occurs in the strain rate range of 0.001 to 1.0 s-1 and causes grain refinement. At a strain of 0.2, the lower strain rate branch of the bifurcation finds superplastic deformation as an attractor when deformed in the strain rate range about 10-5 s 1 (extrapolated as a mirror reflection) to 10-1 s-1, which continues on further straining. The higher strain rate branch of the first bifurcation leads to a DRX attractor only after a strain of about 0.2 (critical strain for DRX). This bifurcation occurs in the strain rate range of 10 to 103 su (the higher strain rate value is an extrapolated one on the basis of a mirror reflection of the domain).

The material exhibits flow instabilities at lower temperatures and higher strain rates as shown by the instability limit in Fig. 30. These instabilities manifest as adiabatic shear bands which are intense at lower temperatures and higher strain rates and flow localization under other conditions. Typical microstructures of specimens deformed at 750°C and at three strain rates in the instability region are shown in Figs. 36 a-c. An intense adiabatic shear band with associated cracking is recorded in the first case and flow localization is more diffused in others. These conditions may be avoided in processing this material.

(iii)design of hot working process On the basis of the constitutive behavior of iron aluminide alloy as revealed in the processing maps, the following hot working schedules may be designed for the bulk working of the material.

Since the material exhibits a change in the mechanism of hot deformation at strains above 0.1, it is beneficial to"condition"the material by hot working it at 1100°C and at strain rates in the range 0.001-1.0 s-l using small strains. This may be done either by forging or rolling. Once the billet is conditioned, the material has extensive workability at temperatures above 1100°C both at higher strain rates (10 s-1) due to DRX and lower strain rates (< 0. 1 s-1) due to superplastic deformation. The higher strain rate domain may be exploited for continuous rolling of the material since this process is generally done at higher speeds.

However, component manufacture from the sheets may be done by superplastic forming associated with diffusion bonding with a suitable material. Manufacture of other forged components are best done in the high strain rate domain using processes including drop forging which is a cost effective process.

Gas Atomized Powders (i) kinetic analysis Within the domain of DRX (in the temperature range 950-1150°C and strain rate range 0.001-1.0 s-1), kinetic analysis using Eq. (1) has been conducted. Considering the variation of log (a) vs. log () in the above limited temperature and strain rate ranges to be approximately linear, the value of the stress exponent is estimated to be about 4.4 (an average strain rate sensitivity of about 0.23). The Arrhenius plot giving the variation of log (a) with (1/T) is shown in Fig. 37, from which an apparent activation energy of about 465 kJ/mole has been estimated for the process of DRX in this material. This value is higher than that for the diffusional processes in FeAl, as is commonly observed for DRX.

It is customary to correlate the grain size variations in the DRX domain with the Zener-Hollomon parameter, Z, given by: eexp [QIRT] Such a variation is shown in Fig. 38, which is linear as expected for the DRX process. Such plots are useful in controlling grain size in the material during processing.

(ii) processing maps The processing maps developed at strains 0. 1, 0.2,0.3,0.4 and 0.5, on the basis of the flow stress data given in Table 3, are shown in Figs. 39-43, respectively. These maps are obtained by a superimposition of the instability maps over the power dissipation maps. In each of the processing maps, the regime of flow instability as predicted by the continuum criterion given by Eq. (3) is delineated by a thick line (marked as"0") running across several efficiency contours belonging to the power dissipation maps. The power dissipation maps show isoefficiency contours, which represent the relative rate of entropy production occurring in the material due to microstructural dissipation. They can also be termed as"microstructural trajectories"since they actually represent the rate of change of microstructure occurring during hot deformation. The curvature of the trajectories changes when the temperature is increased beyond about 875°C, the temperature at which dissolution of fine carbide particles is likely to occur. Since the material undergoing hot deformation is a non-linear dissipator of power, the microstructural trajectories get attracted to basins of lower dissipative energy and form domains where the efficiency of power dissipation is maximized. These domains represent specific microstructural mechanisms, which may be identified by metallographic examination of specimens deformed in this domain.

Referring to the map obtained at the strain of 0.1 (Fig. 39), the single domain observed has a peak efficiency of power dissipation of 44 % occurring at about 1075 °C and 0.1 s n.

Maps obtained at higher strains (Figs. 40-43) are not significantly different from that obtained at a strain of 0.1, although the peak efficiency of the domain referred to above has slightly decreased from 44 % to 40 %. In high stacking fault energy metals, the maximum efficiency of power dissipation for DRX is about 50 % while it is about 35 % in low stacking fault energy metals. Thus the observed value of 40-44 % suggests that this iron aluminide alloy has a medium stacking fault energy.

Typical microstructures obtained on specimens deformed at a strain rate of 0.1 s n and at temperatures of 1050 and 1100°C are shown in Figs. 44 a-b. The microstructures exhibit fine equiaxed grain structure with irregular grain boundaries typical of dynamic recrystallization. The variation of average grain diameter with deformation temperature is shown in Fig. 45, which exhibits a sigmoidal curve typically observed when dynamic recrystallization occurs. The variation of the efficiency of power dissipation with temperature as obtained from the processing map (Fig. 43) at the strain rate of 0.1 s-1 is also shown in Fig. 45. It may be noted that the temperature at which 50 % variation in grain size has occurred coincides with the temperature for the peak efficiency in the domain (called DRX temperature). This is also a typical feature of the DRX process. At deformation temperatures lower than the DRX temperature, the grain size gets finer (10 pm or less) than the initial value (13, mm) while it increases (to about 20, um) at higher temperatures. The microstructures of the material deformed at 1100°C and at strain rates of 0.01 and 1.0 s-l are shown in Figs. 46 a-b and the variation of grain size with strain rate in the domain is plotted in Fig. 47, both of which show that the grain size is finer at higher strain rates and the variation is linear with respect to log strain rate. At the optimum temperature and strain rate for DRX, the workability of the material is maximum. Thus the hot working processes may be designed for workability optimization as well as for microstructural control within this domain of DRX.

As per the instability criterion given by Eq. (3), the material exhibits flow instabilities at strain rates higher than 10 s-1 in the temperature range 950-1100°C. These instabilities manifest as bands of flow localization as seen in the microstructure of the specimen deformed at 100 s-'and 1050'C given in Fig. 48. The conditions of instability predicted by Eq. (3) may be avoided in processing this material.

(iii) desigh of hot workihg pYOCesses On the basis of the constitutive behavior of iron aluminide alloy as revealed in the processing maps, the following hot working schedules may be designed for the bulk working of the material.

Since the material undergoes dynamic recrystallization in the temperature range 950- 1150 °C and strain rate range 0.001-1.0 s-1, with a peak efficiency of power dissipation (44%) occurring at 1075 °C and 0. 1 s-l, these parameters represent the optimum conditions for working this material. Manufacture of components is best done in this low to moderate strain rate domain using processes such as extrusion and press forging.

Conversely, at high strain and high strain rates at temperatures between 950 and 1000 °C, flow instabilities occur, which should be avoided. This suggests that this alloy is not optimized for processing by, for example, continuous rolling, which is generally done at higher speeds, or drop forging, which involves high strain rates.

Additionally, the grain size varies sigmoidally with temperature at the strain rate corresponding to peak efficiency in the DRX domain while the grain size decreases with strain rate in a linear fashion. Therefore, control of both the temperature and strain rate environment is important to obtain the desired microstructure. Further, the selection of this environment may be aided by the use of the Zener-Hollomon parameter.

The difference in the behavior of the gas atomized and water atomized powder compacts is attributed to the reduced oxide particle content in the former case.

Structural and Electrical Resistance Applications The FeAl based alloys and the processing methods developed here are intended for use in industrial and domestic applications. Some examples of possible uses include as heat treatment and furnace fixtures in the thermal processing industry, as heating elements and resistance alloys, and as forged components such as automotive valves. In these applications, the superior corrosion resistance of FeAl based alloys coupled with the reduction in manufacturing costs are attractive.

Although the present invention has been described in connection with preferred embodiments thereof, it will be appreciated by those skilled in the art that additions, deletions, modifications, and substitutions not specifically described may be made without department from the spirit and scope of the invention as defined in the appended claims.