Journal of Materials Science

, Volume 48, Issue 19, pp 6535–6541

Microstructure and mechanical properties of two-phase Fe30Ni20Mn20Al30: part II mechanical properties


  • X. Wu
    • Thayer School of Engineering, Dartmouth College
    • Thayer School of Engineering, Dartmouth College
  • H. Wu
    • Thayer School of Engineering, Dartmouth College
    • State Key Laboratory of Powder MetallurgyCentral South University

DOI: 10.1007/s10853-013-7449-8

Cite this article as:
Wu, X., Baker, I. & Wu, H. J Mater Sci (2013) 48: 6535. doi:10.1007/s10853-013-7449-8


This paper describes the mechanical properties of B2/L21 two-phase Fe30Ni20Mn20Al30 (at.%) in both the as-cast condition and after a 72 h anneal at 823 K. The temperature dependence of the compressive strength of Fe30Ni20Mn20Al30 showed three distinct regions: (1) brittle fracture at low temperature, (2) plastic flow with a rapid decline in yield strength from 1500 to 250 MPa from the brittle-to-ductile transition temperature (BDTT) to 873 K, and (3) a slight decrease in yield strength to ~150 MPa from 873 to 1073 K. Interestingly, the BDTT (573 K) exhibited by the coarser microstructure present in 72 h annealed material was lower than that of the as-cast alloy (623 K). Using both differential scanning calorimetry and in situ heating in a transmission electron microscope, an L21-to-B2 transition was found at 750 ± 25 K. A mixture of intergranular fracture and transgranular cleavage was observed after room temperature compression while only cleavage was found at 673 K. All the specimens deformed extensively without fracture when tested at temperatures higher than 673 K. The strain rate had little effect on the strength at 573 K and a moderate effect at 873 K with a strain-rate sensitivity exponent value of 0.1.


Recently, alloys in the quaternary Fe–Ni–Mn–Al system has been investigated because of their potentially useful mechanical properties. Several b.c.c.-based two-phase alloys, such as Fe30Ni20Mn25Al25 [16] and Fe35Ni15Mn25Al25 [7, 8], have been discovered in this system with the various elemental contents ranging from 15 to 35 at.%. These alloys, which exhibit high strength at room temperature, typically consist of either alternating b.c.c. and B2-ordered phases or alternating B2- and L21-ordered phases with nanoscale phase widths [1, 9]. After long-time aging of these alloys, Mn-rich precipitates nucleate and grow, resulting in a ~50 % increase in hardness [6, 7]. Thus far, the temperature dependence of the yield strength of only one of these alloys, the B2 (ordered b.c.c.)/b.c.c. alloy Fe30Ni20Mn25Al25, has been reported [3].

Strongly ordered intermetallic compounds have aroused considerable interest for their potential as high-temperature structural materials. Incorporating a B2-ordered (ordered b.c.c.) phase into an L21-ordered (Heusler-type) matrix, or vice versa, has been shown to be an effective method for improving mechanical strength [1012]. Strutt and co-workers [10, 11] were the first to study the mechanical properties of a two-phase B2/L21-ordered alloy. They studied Ni50Al35Ti15, which in the as-cast state exhibited a tweed structure. The microstructure of aged Ni50Al35Ti15 was found to consist of semi-coherent B2-ordered NiAl particles in an L21-ordered Ni2AlTi matrix of similar volume fraction and a cube-on-cube crystallographic relationship [10]. The creep strength of the two-phase, polycrystalline NiAl–Ni2AlTi structure was much higher than that of single-phase polycrystalline Ni2AlTi, which itself showed greater creep resistance than single crystal NiAl [10, 11]. The creep rate of the two-phase alloy was comparable to that of the nickel-based superalloy MAR-M200 at similar homologous temperatures [10]. Dislocation analysis of a NiAl–Ni2AlTi specimen crept at 138 MPa at 1173 K found both undissociated a〈110〉 dislocations and anti-phase boundary-coupled a〈100〉 dislocations in the Ni2AlTi phase, where a is the lattice parameter of a b.c.c. unit cell of which the Ni2AlTi unit cell is composed [10, 11].

Several other alloys containing coherent L21-ordered Ni2AlTi particles, which were initially aligned along the 〈100〉 direction either in B2-ordered NiAl or NiTi matrices have been studied in the Ni–Al–Ti system [1214]. Long-time aging led to a loss of coherency and the formation of globular Ni2AlTi particles with 〈100〉 dislocations at their interfaces [1214]. These two-phase alloys displayed significantly improved high temperature strength and better creep properties than either of the constituent phases [12, 13].

A similar microstructure was observed in NiAl–Ni2AlHf alloys with L21-structured Ni2AlHf particles forming coherently on {111} planes of the NiAl matrix with a cube-on-cube relationship [15, 16]. After annealing, the particles grew spherically with a loss of coherency and misfit dislocations were present at the interfaces [15, 16]. The alloys showed an initial increase in hardness followed by a decrease after annealing at 973–1173 K with the optimum anneal producing a hardness of 700 HV [16]. The creep rate was appreciably decreased by the fine precipitation of L21-structured Ni2AlHf phase in the NiAl matrix [16].

Finally, Takeyama and Liu [17] studied NiAl–Ni2AlHf alloys in which the volume fraction of the Ni2AlHf phase was varied from ~15 to 96 %. A lattice misfit of 5 % resulted in no coherency between the B2- and L21-structured phases [17]. The hardness and yield strength increased with increasing volume fraction, while the ductility decreased [17].

In this paper, the mechanical behavior of Fe30Ni20Mn20Al30 was determined at a fixed strain rate as a function of temperature and as a function of strain rate at two elevated temperatures (573 and 873 K) on two different microstructures: the very fine microstructure in the as-cast alloy and the coarser microstructure observed after a 72 h anneal. The fracture surface morphology was also examined. An L21-to-B2 transition was studied using both differential scanning calorimetry (DSC) and in situ heating in a transmission electron microscope (TEM).



Ingots of Fe30Ni20Mn20Al30 were arc-melted in a water-chilled copper crucible under an argon atmosphere from high-purity elemental pieces. An additional 5 wt.% Mn was added to compensate for the loss of Mn by evaporation during melting. The ingots, which were ~5 cm diameter, ~50 g buttons, were flipped and re-melted twice after the initial melting to insure a homogeneous mixture. Some ingots were annealed in air at 823 K for 72 h, followed by air-cooling. This annealing temperature has been shown to be optimum for increasing both the hardness and yield strength of FeNiMnAl alloys [1, 3].

TEM specimens (3 mm in diameter) were first produced by electro-discharge machining, followed by grinding to ~200 μm in the thickness. A Struers Tenupol 5 was used to produce the thin foils using an electrolyte of 20 % nitric acid, 10 % butoxyethanol and 70 % methanol held at ~260 K at a voltage of ~10 V and a current of ~40 mA. After electropolishing, specimens were washed alternately in ethanol, methanol and ethanol, followed by a final rinse in fresh methanol. In situ heating studies at a heating rate of 20 K/min up to 873 K were performed on the resulting thin foils using a Gatan double-tilt heating holder in an FEI Tecnai F20 field emission gun (FEG) TEM operated at 200 kV on a specimen previously given a 72 h anneal.

Thermal analysis

DSC was performed on the as-cast alloy with a TA Q20 DSC under an argon atmosphere. A thin specimen of ~20 mg was sealed in a copper pan, which was covered by a lid and compressed to insure thermal contact with the specimen. An identical empty copper pan was used as a reference. The specimen was heated up to 1000 K at a rate of 20 K/min. Following cooling to room temperature, a second run was performed to provide comparison.

Mechanical testing

Cuboidal (~3 × 3 × 8 mm) compression specimens were machined and polished with 600-grit silica paper before testing. Specimens were tested between superalloy Inconel 718 platens that are usable for high temperature tests with a hydraulic MTS equipped with a furnace. The specimens were pre-loaded to a load of ~100 N before the tests started. The load and displacement of the piston were recorded. Compression tests were performed as a function of temperature at a fixed strain rate (5 × 10−4 s−1) and as a function of strain rate (5 × 10−6–5 × 10−1 s−1) at elevated temperatures (573 and 873 K) on the as-cast alloy and the alloy after a 72 h anneal. At least two tests were performed for each testing condition.

After compressing at room temperature and 673 K, the fracture surfaces of both the as-cast specimen and the specimen after a 72 h anneal were characterized using an FEI XL-30 FEG scanning electron microscope (SEM) operated at 15 kV.


Since the phase widths of the as-cast alloy coarsened by a factor of five, from ~5 to ~25 nm, after a 72 h anneal at 823 K, the temperature dependence of the compressive strength of Fe30Ni20Mn20Al30 was determined in both the as-cast state and after a 72 h anneal at 823 K. The yield strength results are summarized in Fig. 1. The as-cast samples fractured at ~1300 MPa at room temperature before any evidence of yielding. The fracture strength increased as the temperature increased, and a brittle-to-ductile transition occurred at 623 K, at which point a yield strength of ~1500 MPa was observed, see Fig. 1a. The 72 h annealed alloy showed similar behavior to the as-cast alloy but the brittle-to-ductile transition temperature (BDTT) occurred ~50 K lower at 573 K, again with a yield strength of ~1500 MPa, see Fig. 1b, but in this case there was no increase in fracture strength with increasing temperature. Above the BDTT the yield strength for both the as-cast and annealed alloy dropped dramatically with increasing temperature up to 873 K, and then decreased more slowly to ~150 MPa at 1073 K.
Fig. 1

The temperature dependence of the compressive strength of Fe30Ni20Mn20Al30 in a the as-cast state, and b after a 72 h anneal at 823 K. The L21 to B2 transition temperature is indicated

Secondary electron (SE) images of the fracture surfaces after the compression tests are shown in Fig. 2. The as-cast specimens and 72 h annealed specimens exhibited similar fracture surfaces. After compression tests at room temperature, both intergranular fracture and transgranular cleavage fracture modes were observed, while only transgranular cleavage was observed after the compression tests at 673 K. When the testing temperatures were higher than 673 K, all the specimens either buckled or deformed extensively without fracture.
Fig. 2

SE images of the fracture surface of a as-cast and b 72 h annealed specimens compressed at room temperature showing intergranular fracture and transgranular cleavage; c as-cast, and d 72 h annealed specimens compressed at 673 K showing only transgranular cleavage

Thermal stability

The thermal stability of Fe30Ni20Mn20Al30 was examined by performing DSC measurements. DSC curves for the as-cast alloy are shown in Fig. 3. An endothermic peak occurred at ~725 K in the heating curve, with a small enthalpy of transformation of 0.17 kJ/mol, determined by integrating the area under the peak. Upon cooling an exothermic peak was observed at ~775 K in the cooling curve.
Fig. 3

DSC curves a heating at 20 K/min and b cooling curves for as-cast Fe30Ni20Mn20Al30. A peak indicates an endothermic process and a trough indicates an exothermic process

In order to determine the origin of the peaks observed in the DSC measurements, a TEM in situ heating experiment was performed on the 72 h annealed specimen. Figure 4 shows [011] zone axis (ZA) selected area diffraction patterns (SADPs) acquired at different temperatures. L21 superlattice reflections were clearly observed at 573 K, as indicated in Fig. 4a. The intensity of the L21 spots started to decrease when the temperature reached 673 K. When the temperature was above 723 K, the L21 spots were no longer observable. The B2 superlattice reflections did not change during the heating experiment. The transformation temperature obtained from TEM in situ observations broadly agrees with the DSC results. However, for the TEM observations exactly when the L21 diffraction spots are no longer visible is a matter of judgment, which clearly gives rise to some error. Figure 5 shows [011] convergent beam electron diffraction (CBED) patterns acquired at 773 K from the [011] ZA showing that both phases present are B2 structured. Thus, the in situ heating experiment indicated that the peaks in the DSC curves are due to an L21-to-B2 transformation upon heating and the B2-to-L21 transformation upon cooling. A significant decrease in L21 order initiated at ~673 K and the L21 reflections were no longer visible by 723 K.
Fig. 4

SADPs taken at [001] ZA at a 573 K, b 673 K and c 723 K, respectively. The L21 superlattice spots are highlighted. The dashed rectangles indicate L21 systematic rows
Fig. 5

CBED patterns obtained at 723 K from the individual phases of 72 h annealed Fe30Ni20Al20Mn30 showing that both exhibit a B2 structure. Note the absence of L21 reflections

Since, irrespective of whether the alloy was annealed or not, Fe30Ni20Mn20Al30 was brittle at room temperature, the effect of strain rate on the yield strength was studied at 873 K for the as-cast condition and 573 K for the 72 h annealed condition, see Fig. 6. The results show that at 573 K the yield strength was only weakly strain-rate dependent, but at 873 K the strain rate dependence was much more marked.
Fig. 6

Strain-rate dependence of a 72 h annealed Fe30Ni20Mn20Al30 at 573 K, and b as-cast Fe30Ni20Mn20Al30 at 873 K


The L21–B2 transition of Fe30Ni20Mn20Al30 occurs at 750 ± 25 K. This temperature is lower than the L21–B2 transition temperatures of L21-structured Fe55Mn19Al26 (~890 K) [18] and Fe2AlMn single crystal (~898 K) [19]. The estimated enthalpy of the transformation for Fe30Ni20Mn20Al30 is ~0.17 kJ/mol. This value is about 80 % less than that for Fe55Mn19Al26 (~0.94 kJ/mol) [18] and Fe2AlMn (~1.1 kJ/mol) [19]. There are two possible reasons for the lower transformation temperature and smaller transformation enthalpy for Fe30Ni20Mn20Al30. First, the microstructure of Fe30Ni20Mn20Al30 consists of both B2 and L21 phases with a volume fraction of L21 of ~50 %, whereas Fe55Mn19Al26 and Fe2AlMn are L21-structured single-phase alloys. Second, the composition of the L21 phase in Fe30Ni20Mn20Al30 is Fe–39Ni–12Mn–34Al, indicating that the positions of the atoms are not strictly L21 (one type of atom occupies all the corner sites, and the other two types of atoms sit in the body center sites alternately). Thus, the L21 phase in Fe30Ni20Mn20Al30 is less “perfectly” ordered compared with the L21 phase in Fe2AlMn or in Fe55Mn19Al26.

The temperature dependence of the yield strength of Fe30Ni20Mn20Al30 shows three distinct regions, see Fig. 8:
  1. (1)

    when the temperature was below the BDTT, specimens fractured before yielding with no clear trend of increasing fracture strength with increasing the temperature;

  2. (2)

    when the temperature was between the BDTT and 873 K, a rapid decline in yield strength from ~1500 to ~250 MPa was observed;

  3. (3)

    above 873 K, the strength decreased more slowly to ~150 MPa and diffusional processes may have contributed to the deformation.


Apart from the first region, the other two regions are broadly similar to the behavior observed in B2 compounds. For example, at intermediate temperatures (~0.4Tm < T < ~0.6Tm), the yield strength of FeAl [20] and NiAl [21] decreases rapidly with increasing temperature, and above ~0.6Tm, a slight decrease is observed in the yield strength of NiAl due to diffusion-assisted processes [21]. It is worth noting that the elevated temperature strength of the B2/b.c.c. alloy Fe30Ni20Mn25Al25 experienced a dramatic drop from 1400 to 420 MPa between 673 and 773 K, followed by a slight decrease to 200 MPa at 1073 K [3], which is similar to the behavior of Fe30Ni20Mn20Al30 studied here.

Interestingly, the BDTT 623 K for the as-cast alloy and 573 K for the 72 h annealed alloy occurs at ~0.4Tm. Both are slightly below the L21–B2 transition temperature of 723 K. The BDTT of B2-structured NiAl is also around ~0.4Tm, which corresponds to the onset of thermally activated deformation mechanisms [22, 23]. In post-mortem TEM analysis of Fe30Ni20Mn20Al30, only a 〈100〉 dislocations were observed [24], which provide only three independent slip systems. Thus, only limited ductility would be expected since five independent slip systems are required for general polycrystalline plasticity [25]. When the temperature is above the BDTT, it has been suggested that dislocation climb can accommodate imposed strains in B2 compounds, hence increasing the ductility at higher temperatures [26]. It is interesting to note that in a separate study, the 72 h annealed material was found to show significantly improved wear behavior compared to the as-cast alloy, a feature that may again be related to the improved ductility of the annealed material [27].

Both as-cast and 72 h annealed Fe30Ni20Mn20Al30 showed similar fracture modes. At room temperature, a mixture of intergranular fracture and transgranular cleavage was observed. At 673 K, only transgranular cleavage occurred. A transition from intergranular fracture to transgranular cleavage with increasing temperature is well known [28]. When intergranular fracture occurs, the grain boundary is weaker than the slip plane, whereas for transgranular fracture, the slip plane is the weaker component [29]. This fracture mode change was also observed in B2-structured FeAl compounds [20, 30]. Polycrystalline NiAl generally fractures intergranularly at room temperature [31] and a transition to transgranular cleavage has been reported at 673–773 K [31, 32].

As shown in Fig. 6, strain rate has little effect on the strength for the 72 h annealed Fe30Ni20Mn20Al30 at 573 K, when it just starts to yield before fracture, but has an effect for the as-cast Fe30Ni20Mn20Al30 at 873 K, when the yield strength drops to ~250 MPa. The yield strength shows strain-rate hardening behavior, i.e., the faster the strain rate, the higher the yield strength. Figure 7 indicates that the relationship between the yield stress and the strain rate is
$$ \sigma = C \cdot \varepsilon^{0.1} , $$
where C is a constant and the strain-rate sensitivity exponent has a value of 0.1. In other words, the stress exponent m in the power law relation
$$ \dot{\varepsilon } = K\sigma^{m} , $$
is 10. The same stress exponent value was reported by Yang and Dodd for polycrystalline NiAl [33] and is a typical value for dislocation glide in b.c.c. metals [34].
Fig. 7

Strain-rate dependence of as-cast Fe30Ni20Mn20Al30 at 873 K showing that the strain-rate sensitivity is ~0.1

Figure 8 is a schematic summarizing the strength as a function of temperature for the two alloy conditions with the effect of strain rate indicated.
Fig. 8

Schematic diagram showing three regions of the strength versus temperature curve of Fe30Ni20Mn20Al30 in both the as-cast and 72 h annealed condition. The effect of strain rate at 873 K is also indicated


An investigation was conducted on the mechanical properties of Fe30Ni20Mn20Al30. The results of these investigations are summarized as follows:
  1. (1)

    A L21–B2 transformation was observed using both in situ heating TEM and DSC to occur at 750 ± 25 K.

  2. (2)

    The temperature dependence of the strength of Fe30Ni20Mn20Al30 showed three distinct regions: (1) fracture before showing any yield below the BDTT, (2) a rapid decline in strength from ~1500 to ~250 MPa between the BDTT and 873 K and (3) a slower decrease to ~150 MPa above 873 K.

  3. (3)

    Fe30Ni20Mn20Al30 shows brittle fracture behavior at room temperature. After a 72 h anneal at 823 K, the alloy exhibited a lower BDTT than the as-cast alloy.

  4. (4)

    Both intergranular fracture and transgranular cleavage were observed after room temperature compression while only transgranular cleavage was found after compression at 673 K.

  5. (5)

    Strain rate has almost no effect on the strength of Fe30Ni20Mn20Al30 at 573 K and has moderate effect at 873 K with a strain-rate sensitivity m value of 0.1.



This research was supported by the US Department of Energy (DOE), Office of Basic Energy Sciences (DOE Grant DE-FG02-07ER46392). The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing official policies, either expressed or implied of the DOE or the U.S. Government.

Copyright information

© Springer Science+Business Media New York 2013