THERMAL EXPANSION DATA ON SEVERAL IRON- AND NICKEL-ALUMINIDE ALLOYS W. D. Porter and P. J. Maziasz Metals and Ceramics Division, Oak Ridge National

THERMAL EXPANSION DATA ON SEVERAL IRON- AND NICKEL-ALUMINIDE ALLOYS W. D. Porter and P. J. Maziasz

Metals and Ceramics Division, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, T N 37831

(Received June 7, 1993) (Revised July 15, 1993)

Introduction

Iron and nickel aluminides are ordered intermetallic alloys that are being developed for high-temperature structural applications because they have good oxidation resistance at >1000°C due to the formation of adherent alumina films. More complex aluminide alloys (Fe3A1, Ni3A1, and FeAI types) have recently been developed to improve their mechanical properties behavior, including the combination of room-temperature ductility and high- temperature strength, and to improve oxidation/corrosion resistance [I-7]. The Fe3AI alloy composition has a B2-ordered phase structure at about 550 to 900°C but transforms to the DO 3 ordered phase at lower temperatures.

Fe3Al-type alloys have been developed around a hyperstoichiometric Fe-28A1-5Cr (at. %) composition with smaller additions of Nb and C [1-3]. The FeAI alloy system has the ordered B2 phase structure from room temperature to

>1100°C, and alloys are being developed from an Fe-36A1 (at. %) base composition [4,5]. The Ni3AI alloy exhibits the L12 ordered crystal structure from room temperature to >1300°C, and alloys are currently being developed near a hypostoichiometric Ni-16AI-8CY (at. %) composition, with varying minor additions of B, Zr, and Mo [6,7].

While a considerable amount of mechanical property, structural, microstructural, and corrosion data has been generated in the development of the iron- and nickel-aluminide alloys, there have been correspondingly little physical properties data. Physical properties like thermal expansion are basic material parameters used by engineers to select materials and by designers to perform stress analyses. The purpose of this paper is to present the preliminary thermal expansion data measured on leading or representative new alloys being developed from each of these three iron- or nickel-aluminide alloy types.

Three representative iron and nickel aluminides were used in this study. They were FA-129 (an Fe3A1 alloy), FA-385 (an FeA1 alloy), and IC-221M (an Ni3A1 alloy). The nominal composition of each of the alloys used is shown in Table 1. The FA-129 used was fabricated from material that had been electroslag remelted (Special Metals Corporation) then extruded at 1000°C (Oak Ridge National Laboratory) using an 8:1 reduction ratio [3]. The sample used for expansion measurements was electrodischarge machined (EDM) from the transverse direction of the extrusion. The sample had a diameter and length of 4 and 25 mm, respectively, and was tested in the as- extruded condition. The FA-385 sample had the same dimensions as the FA-129 sample and was EDM-cut from the top of an ingot that had been arc-melted and drop-cast into a water-chilled copper mold. The IC-221M sample was machined from the gauge section of an investment-cast fatigue test bar and had a diameter of 6 mm and a length of 25 mm. The FA-385 and IC-221M samples were tested in the as-cast condition.

TABLE 1

Nominal Allo), Composition (at. %)

Alloy A1 Cr Mo Zr Nb C B Fe Ni

FA-129 28.0 5.0 0.5 0.2 Balance

FA-385 35.8 0.2 0.05 0.13 Balance

IC-221M 15.9 8.0 0.8 1.0 0.04 ^Balance

1043

0 9 5 6 - 7 1 6 X / 9 3 $6.00 + .00

Expansion measurements were made using a dual push-rod differential dilatometer (Them Industries). The advantages of such an instrument have been reported by Plummer [8]. The dilatometer has a horizontal configuration with push rods and sample holder constructed of high-density alumina. A sapphire rod with a length of 25 mm was used as the reference material. Differential changes in length between the sample and reference are transmitted to a linear variable differential transformer (LVDT) that is mounted on an Invar rod using separate pairs of leaf springs for the coil and core. The leaf springs provide frictionless movement of the LVDT and also maintain a push rod load of 25 to 30 gm on the sample and reference. The LVDT is housed in an enclosure that is maintained at 40°C by means of water circulated from a constant-temperature bath. A platinurn/platinum-10% rhodium (type S) thermoeouple was used to monitor the sample temperature. The accuracy of the expansion measurements was determined to be better than +9% over the range of 20 to 1450°C by means of verification tests using sapphire and tungsten reference materials in the sample and reference positions.

Oxidation of the samples during the dilatometer runs was minimized by evacuating the system using a mechanical vacuum pump followed by backfilling with titanium-gettered helium. This process was repeated three times prior to each dilatometer run. During the test, a helium flow rate of 5 ml/min was maintained at a slight overpressure of 20.7 kPa (0.204 atm).

Expansion measurements were made using a computerized data acquisition system while the samples were heated and cooled at a rate of 3°C/min. Isothermal holds were maintained for 20 min at one or two temperatures during the heating and cooling ramps as well as at the maximum temperature for the test. These holds are intended to minimize any temperature gradients developed during the ramps. FA-129 and FA-385 were tested to 1300°C and IC-221M was tested to 1100°C. Data sets were stored at 30-s intervals.

Results

The linear thermal expansions of FA-129 and FA-385 are plotted as a function of temperature in Fig. 1 along with literature values for type 316 austenitic stainless steel and FeAI [9,10]. Also shown in Fig. 1 are the best fits to the experimental data determined by the method of least squares. The equations for the fitted curves are:

FA-129 AO/'o

FA-385 AQ/o

= - 0.02653 + 0.001436T + 4.403 x 10"TT 2 + 9.596 x 10-10T 3 (20 < T < 543"C);

= - 0.4969 + 0.003791T - 2.736 x 10"6T 2 + 1.794 x 10-9T 3 (543 < T < 924"C);

= - 4.488 + 0.01438T -1.090 x 10-ST 2 + 3.295 x 10-9T 3 (924 < T < 1300"C);

= - 0.061450 + 0.002119T - 3.721 x 10-TT + 6.542 x 10-10T 3 (20 < T < 1190"C);

= - 367.23 + 0.8858T - 7.085 x 10"4T 2 + 1.896 x 10-TT 3 (1190 < T < 1300"C);

where AQ/ois the linear thermal expansion (%), and T is the temperature in "C. The regression coefficients for the equations are 0.99999, 0.99999, and 0.99997 for FA-129 and 0.99996 and 0.99993 for FA-385, respectively.

70 I I t I I I

FIG. 1.

e0

4 0

i-

10 0

.--o-- FA.~ (Fq~U) ~cP

--~-- F*-3eS(F~U) riP" ~o

• 8S316 (Flog. 0) ~ ~ : t C ~ ee

• Fe~(p~. lo) ~ e e ee

~ ° I l

400 800 800 1000 1200 1400

T E M P E R A T U R E ('(3)

A comparison of the thermal expansion of FA-129 and FA-385 with type 316 stainless steel (SS316) and binary FeA1.

The temperature limits chosen for the curve fitting are those indicated by sharp peaks in the instantaneous coefficient of expansion as shown in Fig. 2(a) and 2(b). The phase transition temperatures associated with these peaks are 543"C for the D03 ~ B2 and 924"C for the B2 ~ disordered tx phase transitions in Fe3A1 alloy (FA-129), and 1190"C for B2 ~ disordered ~ phase transition in the FeA1 alloy (FA-385). The instantaneous coefficient of expansion was determined by first smoothing both the temperature and expansion data using an unweighted 5-point moving average. An incremental slope method utilizing A expansion/A T from the current and following data pair was then used to calculate the first derivative of the smoothed expansion data with respect to temperature. Values for the mean coefficient of expansion of FA-129 and FA-385 are shown in Table 2.

70

8O

O

< 20 _z

10

i i

--,O--- FA-129 (FeAI)

ORM. OWG g3M 7S93

I I

(a)

924*C

(Fe3AI) I (FeAI) (a Fe)

B

I I I I I I

200 400 600 800 1000 1200

TEMPERATURE (*C)

70

6O

5O

401

3O

2O

10

0 t400

O~-D~4~ 13M-71n12

t I I I I I

(b) FA-385 (FeAt)

1190~C

I I I I I I

200 400 600 800 1000 1200 1400

TEMPERATURE (*C)

FIG. 2. Instantaneous coefficient of thermal expansion of (a) FA-129 and (b) FA-385 as a function of temperature. DO 3 ~ B2 and B2 ---) disordered ct-Fe transitions appear as sharp peaks.

TABLE 2

Mean Coefficient of Expansion of Iron and Nickel Aluminides (ppm/*C)

Temperature (°C) FA-129 FA-385 IC-221M

100 15.4 18.3 12.8

200 15.9 19.0 13.1

300 16.7 19.6 13.4

400 17.9 20.3 13.7

500 19.2 20.6 14.0

600 20.4 21.1 14.3

700 21.0 21.6 14.7

800 21.8 22.2 15.2

900 22.8 22.9 15.8

1000 23.4 23.8 16.6

1100 23.4 24.8

1200 23.5 26.5

The linear thermal expansion of IC-221M is plotted as a function of temperature in Fig. 3 along with values from the literature for IN 713C [11]. The best fit to the experimental data for IC-221M is also shown in Fig. 3 and has the equation:

IC-221M

AO/o

The regression coefficient for the equation is 0.99998. The instantaneous coefficient of expansion of IC-221M is shown in Fig. 4 as a function of temperature. The mean coefficient of expansion of IC-221M is presented in Table 2.

2.0

1200 1.5

<1 1.0

¢/) Z

0.5

ORNL,,OWO I~ I - ~ I

I I I I I

7

200 400 600 800 1000

TEMPERATURE (*C)

30

FIG. 3. A comparison of the thermal expansion of IC-221M and IN 713C.

ORNL-DWG II3M-/U

I I I I I

0

~

Z

z 14

10 0

I I I 1 I

200 400 600 800 1000

TEMPERATURE (°C)

1200

FIG. 4. Instantaneous coefficient of thermal expansion of IC-221M as a function of temperature. No phase transitions are apparent in the temperature range investigated.

Discussion

Very little information is available in the literature on the thermal expansion of Fe3A1, FeAI, and Ni3Al aluminide alloys, and the information that is available is typically for binary alloys. (There have been some systematic physical properties data presented for NiA1 alloys [12]). Therefore, we will compare our results on the aluminides not only with the limited binary alloy aluminide data, but also with data from relevant commercial high- temperature steels and superaUoys. In the case of the iron aluminides, the phase transition temperatures revealed from analysis of the expansion data are compared with existing binary Fe-A1 phase diagram data.

The thermal expansions of both the FA-129 (Fe3AI) and the FA-385 (FeA1) iron-aluminide alloys are comparable to that found in type 316 stainless steel from room temperature up to about 600°C (see Fig. 1). Above 600"C, the iron-aluminide alloys tend to have a larger expansion than type 316 stainless steel. Above about 1100"C, the FeA1 alloy shows measurably more expansion than the Fe3A1 alloy. Figure I also shows the expansion data of Clark and Whittenberger [10] for binary FeA1 alloys with 41.7 to 48.7 at. % AI (extruded material prepared from mixed prealloyed powders). The expansion values for our FeA1 alloy are slightly higher than predicted by the equation fit of Clark and Whittenberger; however, these authors grouped the data for all eight of their alloys with different compositions together because they concluded that expansivity was not a strong function of composition. If we compare our data on the Fe-35.8AI alloy with Clark and Whittenberger's data (Fig. 5 of ref.

10) on the Fe-41.7A1 alloy, the agreement is quite good. The data on our Fe-35.8 AI alloy follows the trend of their individual data for FeA1 alloys, which showed increased expansion with decreased aluminum content.

Inflection points that are apparent in the expansion data in Fig. 1 for the Fe3A1 and FeA1 alloys can be seen more clearly in the plot of instantaneous coefficient of thermal expansion (CTE) versus temperature in Fig. 2. The peaks correspond to the D03-to-B2 and the B2-to-disordered et phase transitions, respectively, with increasing temperature. The value of 543"C found for the D03-to-B2 phase transition of the Fe3A1 (FA- 129) alloy in this work agrees fairly well with the value of 530"C determined recently for the same alloy composition via high temperature X-ray diffraction measurements [13]. The binary Fe-A1 phase diagram shows this phase transition to occur at 550"C and the B2-to-0~ transition to occur at about 900"C for an Fe-28A1 alloy [1]. McQueen and Kuczynski [14]

found an ordered D03-to-B2 phase transition temperature of 520"C from expansion measurements of a binary Fe-27AI alloy, which is lower than both the phase diagram value and our Fe3A1 data. However, since recent work has shown that both alloy composition and processing-induced microstructure can affect the D03-to-B2 phase transition temperature and kinetics [13], it is difficult to compare the data more closely without knowing the significance of these parameters. The value of 1190"C for the ordered B2-to-disordered - 0~ phase transition measured in the FeAI alloy in this work (FA-385, Fig. 2) is in reasonably good agreement with the value found on the binary phase diagram for an Fe-36A1 alloy [1].

The thermal expansion of the Ni3A1 alloy (IC-221M) presented here was found to be very similar to published data for the commercial nickel-based superalloy 713C (Inconel 713C) [11], as shown in Fig. 3. The close agreement in thermal expansion behavior between the Ni3AI alloy and alloy 713C is consistent with their similar compositions. Comparison of the Ni3AI with the Fe3AI and FeAI alloys (Figs. 1 and 3) indicates that the expansion of the two iron aiuminides is about twice that of the nickel aluminide at all temperatures. The instantaneous CTE data for the Ni3A1 alloy (see Fig. 4) do not indicate any phase transitions in the temperature range studied.

Future work will include thermal expansion measurements of NiA1 and other intermetallic alloys as well as measuring other physical properties (i.e., specific heat, thermal diffusivity, thermal conductivity, etc.) as (unctions of temperature for the aluminides examined here.

Conclusions

The thermal expansion of an Fe3A1 (Fe-28A1-5Cr) alloy and an FeA1 (Fe-36A1) alloy were found to be similar but slightly higher than type 316 austenitic stainless steel from room temperature to near the melting point.

Over this temperature range, the FeA1 alloy showed slightly more expansion than the Fe3AI alloy. The ordered phase transition temperatures determined from analysis of the instantaneous CTE data were in good agreement with the binary Fe-AI phase diagram.

The thermal expansion of the Ni3A1 (Ni- 16Al-SCr) alloy was similar to that of the nickel-based superalloy IN 713C and about one-half that found for the two iron-aluminide alloys.

Ackrl0wledm-aents

The authors thank J. A. Cook for technical assistance during testing and C. G. McKamey and R. K. Williams for review of the manuscript. We also thank B. E. Mercer for preparation of the manuscript and K. Spence for editing. Work sponsored by the U.S. Department of Energy, Assistant Secretary for Conservation and Renewable Energy, Office of Industrial Technologies, Advanced Industrial Concepts (AIC) Division, AIC Materials Program, and by the Office of Transportation Technologies as part of the High Temperature Materials Laboratory User Program under contract DE-AC05-84OR21400 with Martin Marietta Energy Systems, Inc.

References

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

2. C.G. McKamey, C. T. Liu, S. A. David, J. A. Horton, D. H. Pierce, and J. J. Campbell, Development of Iron Aluminidesfor Coal Conversion Systems, ORNL/TM-10793, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., July 1988.

3. V. K. Sikka, C. G. McKamey, C. R. Howell, and R. H. Baldwin, Properties of Large Heats of Fe3Al-Based Alloys, ORNIdTM-11796, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., March 1991.

4. C.T. Liu, V. K. Sikka, and C. G. McKamey, Alloy Development of PeAl Aluminide Alloys for Structural Use in Corrosive Environments, ORNL FPO-90/119, Martin Marietta Energy Systems, Inc., Oak Ridge Natl.

Lab., September 1990.

5. P. F. Tortorelli and P. S. Bishop, Influence of Compositional Modifications on the Corrosion of Iron Aluminides by Molten Nitrate Salts, ORNL/TM-11598, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., January 1991.

6. C.T. Liu and V. K. Sikka, J.. Met. 38, 19 (May 1986).

7. C. T. Liu, V. K. Sikka, J, A. Horton, and E. H. Lee, Alloy Development and Mechanical Properties of Nickel Aluminide (Ni3Al) Alloys, ORNL-6483, Martin Marietta Energy Systems, Inc., Oak Ridge Natl. Lab., August 1988.

8. W.A. Plummet, "Differential Dilatometry - A Powerful Tool," p. 17 in Intlernational Symposium on Thermal Expansion of Solids, American Institute of Physics Conference Proceedings, New York (1973).

9. Y.S. Touloukian, R. K. Kirby, R. E. Taylor, and P. D. Desai, "Thermal Expansion - Metallic Elements and Alloys," p. 1151 in Thermophysical Properties of Matter, Vol. 12, IFI/Plenum Co., New York, (1977).

10. R.W. Clark and J. D. Whittenberger, "Thermal Expansion of Binary CoAl, FeAI, and NiA1 Alloys," p. 189 in Thermal F_,xpansion-8, ed. T. A. Hahn, Plenum Press, New York (1984).

11. Y.S. Touloukian, R. K. Kirby, R. E. Taylor, and P. D. Desai, "Thermal Expansion - Metallic Elements and Alloys," p. 1223 in Thermophysical Properties of Matter, Vol. 12, IFI/Plenum Co., New York, (1977).

12. R.D. Noebe, R. R. Bowman, and M. V. Nathal, Review of the Physical and Mechanical Properties and Potential Applications of the B2 Compound NiAl, NASA Technical Memorandum 105598, NASA-Lewis Research Center, Cleveland, Ohio, April 1992.

13. P. J. Maziasz, C. G. McKamey, O. B. Cavin and T. Zacharia, "Some Effects of Composition and Microstructure on the B2 to D03 Ordered Phase Transition in Fe3A1 Alloys," p. 209 in MRS Proc.

High-Temperature Ordered lntermetallic Alloys V, Vol. 288, ed. I. Baker, R. Darolia, J. D. Whittenberger, and M. H. Yoo, Materials Research Society, Pittsburgh, 1993.

14. H.J. McQueen and G. C. Kuczynski, Trans. Metall. Soc. AIME 215, 619 (1959).