Microstructures and Thermoelectric Properties of Heusler Fe₂VAl Alloys Containing Oxide Nanoparticles

Abstract

The Heusler-type Fe₂VAl alloy is a promising candidate for use in fabricating a thermoelectric power generation device because of its large Seebeck coefficient and high electrical conductivity. However, the high thermal conductivity of this alloy, as a thermoelectric material, degrades its power generation capacity. In this study, to reduce its thermal conductivity, the microstructure of a sintered Fe₂V_1.08Al_0.92 alloy prepared via a powder metallurgical process was modified by adding oxide nanoparticles. Via the dispersion of Al₂O₃ nanoparticles, a sintered Fe₂V_1.08Al_0.92 alloy with fine grains of approximately 200 nm in size was obtained due to the pinning effect on grain growth during sintering. The thermal conductivity was reduced from 16 to 11 W/mK. Upon La₂O₃ addition, the grain size of the Fe₂V_1.08Al_0.92 alloy was reduced to approximately 100 nm and the thermal conductivity was further reduced to 10 W/mK. The difference in grain refinement could be caused by the lower stability of La₂O₃, which facilitated dispersion during ball milling, compared to that of Al₂O₃. As these microstructure refinements negatively affected the electronic properties, the thermoelectric performance of the Fe₂V_1.08Al_0.92 alloy could not be enhanced. However, partial microstructure refinement with sparsely distributed La₂O₃ could slightly enhance the thermoelectric performance due to an appreciable reduction in the thermal conductivity without a considerable degradation in the electronic properties. By using these thermoelectric properties, a simple estimation of thermoelectric power generation, assuming a thermal resistance between the heat sources and thermoelectric module, was conducted. Remarkably, the results suggested that the reduction in thermal conductivity could enhance the output power density and conversion efficiency and reduce the optimal leg length. Thus, practically, controlling the balance between the electronic and thermal properties via microstructural modification is favorable in improving the practicability of the Fe₂VAl alloy by enhancing the power generation capacity and reducing the sizes and masses of thermoelectric devices.

1. Introduction

Thermoelectric energy conversion is attracting increasing attention as a promising technology for use in efficient thermal management and the effective use of energy resources. The practical application of thermoelectric power generation using vast amounts of low-temperature waste heat, in particular, is anticipated. To this end, enhancing the conversion efficiency and reducing the cost of thermoelectric power generation are essential. The performance of a thermoelectric energy conversion system depends strongly on the performances of the thermoelectric materials, which are evaluated using thermoelectric figures of merit ZT = S²σ/κ, where S, σ, κ, and T are the Seebeck coefficient, electrical and thermal conductivities, and absolute temperature, respectively. Therefore, investigating materials with high ZT values is the first step toward the practical application of efficient thermoelectric power generation.

The Heusler-type Fe₂VAl alloy is a promising material for use in thermoelectric power generation at approximately room temperature because of its large thermoelectric power factor PF (= S²σ) [1–3], e.g., 6.8 mW/mK² at 300 K. Additionally, its high mechanical strength, e.g., a bending strength of 800 MPa when sintered, and excellent chemical stability, e.g., a stable oxidation resistance at approximately 850 K in air, may aid in fabricating a durable thermoelectric module [4], which may be used in harsh environments. Moreover, the abundant, low-cost constituent elements of the Fe₂VAl alloy are advantageous in producing low-cost thermoelectric devices on a large scale. The efficient production of the bulk sintered Fe₂VAl alloy has been investigated for practical application [5–7]. However, a drawback of the Fe₂VAl alloy is its high κ as a thermoelectric material, e.g., 28 W/mK [1], resulting in a poor ZT and low energy conversion efficiency in thermoelectric power generation. In addition, the high κ, which necessitates an extended distance between the heat source and sink to yield a large temperature difference in thermoelectric power generation, increases the lengths of the legs of the thermoelectric module. Consequently, the high κ increases the volumes and masses of thermoelectric devices, hindering the practical application of thermoelectric power generation systems. Therefore, reducing κ is critical in the practical application of the Fe₂VAl alloy in thermoelectric power generation.

To reduce the κ of the Fe₂VAl alloy, the introduction of phonon scattering sources, e.g., the lattice distortion induced by heavy element doping [8–11], the antisite defects induced by compositional control [1, 2, 12], and the grain boundaries formed via microstructure refinement [13–17], has been examined. Off-stoichiometric compositional control, which forms lattice point defects, is an excellent method of suppressing the κ of the Fe₂VAl alloy because it may also enhance S by modifying the electronic band structure [1–3, 11], e.g., the κ of the off-stoichiometric Fe₂V_1.08Al_0.92 alloy prepared using arc melting was reduced to 16 W/mK [1].

In this study, to further reduce the κ and improve the thermoelectric performance of the Heusler Fe₂VAl alloy, microstructure refinement using a powder metallurgical method was examined. The reported cumulative lattice component of the κ of the Fe₂VAl alloy indicated that most of the lengths of the phonon mean free paths were in the range 20–300 nm [18], suggesting that nanometer-scale microstructure refinement is effective in reducing the κ of the Fe₂VAl alloy. However, in our previous study, the grains in the alloy powder that were ground to nanometer sizes via ball milling grew to submicrometer sizes during sintering, even though grain growth was suppressed using pulsed electric current sintering (PECS) with rapid temperature control under uniaxial pressure [13]. Therefore, introducing an additional factor to suppress grain growth is required to reduce the grain sizes to the nanometer scale. To this end, the addition of oxide nanoparticles, which inhibit grain growth during sintering, to the Fe₂V_1.08Al_0.92 alloy was examined in this study. Referring to the Zener equation [19–22], which deals with grain growth in materials containing particles, nanometer-scale grain size reduction within the sintered Fe₂VAl alloy could be realized via the dispersion of oxide nanoparticles with sizes of approximately 10 nm. In this study, the effects of oxide nanoparticle addition on the microstructure and thermoelectric properties of the sintered Heusler Fe₂VAl alloy were investigated. In addition, the benefits of the modification of the thermoelectric properties via microstructure refinement were also evaluated in terms of the usefulness of this alloy for application in thermoelectric power generation.

2. Experimental Procedure

In this study, a well-ordered Heusler Fe₂VAl alloy comprising millimeter-sized coarse grains was initially prepared via arc melting and annealing, and the alloy was then ground using high-energy ball milling to reduce the grain size. The obtained Fe₂VAl powder comprising particles composed of nanometer-sized crystallites was sintered via PECS. To suppress the coarsening of the crystallites of the Fe₂VAl phase during sintering, the dispersion of nanometer-sized oxide particles, which could induce the pinning effect on the grain boundaries of the Fe₂VAl phase during sintering, was examined. The effects of oxide particle addition on the crystal phase, microstructure, and thermoelectric properties of the Fe₂VAl alloy were evaluated. The detailed experimental procedures are as follows.

Sintered Heusler Fe₂VAl alloys containing oxide nanoparticles were prepared using a powder metallurgical process. Ingots of the off-stoichiometric Heusler Fe₂V_(1+x)Al_(1−x) (x = 0.08) alloy were prepared via arc melting using Fe (99.99%), V (99.9%), and Al (99.99%) as raw materials. After annealing at 1273 K for 48 h, the ingots were ground to sizes of <100 µm. The obtained Fe₂V_1.08Al_0.92 powder was mixed with 2.5 vol% of oxide powder as an additive and further ground using planetary ball milling. Commercially available La₂O₃ (Kanto Chemical, Tokyo, Japan) and Al₂O₃ powders (Sigma-Aldrich, St. Louis, MO, USA), with respective mean particle sizes of <15 and <13 nm, were used as the additives. For comparison, commercially available La₂O₃ (Kojundo Chemical Laboratory, Sakado, Japan) and Al₂O₃ powders (Kojundo Chemical Laboratory), with respective mean particle sizes of 5 and >53 µm, were also used as the additives. Ball milling was performed for 10 h at 150 rpm using an Ar-gas-filled 500 mL Cr steel pot containing 400 g of Cr steel balls with diameters of 5 mm for 10 g of the powdered mixture. Upon addition of the La₂O₃ powder comprising nanometer-sized particles, another powdered mixture was ball milled for 20 h. The prepared powders were sintered using PECS with a graphite mold at 1373 K for 1 min under vacuum and a uniaxial pressure of 50 MPa. The obtained sintered samples, with typical respective diameters and thicknesses of 10 and 2 mm, were used in measuring κ and then cut into bar shapes with typical sizes of 2 × 2 × 9 mm for use in measuring σ and S. The sintered sample prepared using the powder that was ball milled for 10 h without an additive is denoted FVA. The sintered samples prepared using the powders that were ball milled for 10 h with the Al₂O₃ powders comprising micrometer- and nanometer-sized particles or the La₂O₃ powders comprising micrometer- and nanometer-sized particles are denoted mAO, nAO, mLO, and nLO-10, respectively. In addition, the sintered sample prepared using the powder that was ball milled for 20 h with the La₂O₃ powder comprising nanometer-sized particles is denoted nLO-20.

Crystallographic phase analysis was performed using X-ray diffraction (XRD) with Cu Kα radiation, and the microstructure was observed using scanning electron microscopy (SEM). κ was estimated using the density (D), thermal diffusivity (α), and specific heat (C_p) using the relationship κ = D × α × C_p. D was calculated using the mass and volume of the specimen, and α and C_p were measured in an Ar atmosphere using the laser flash method. S and σ were simultaneously measured in a He atmosphere using a σ and S measurement system (ZEM-3, ADVANCE RIKO, Yokohama, Japan). σ was evaluated using the conventional four-probe direct-current technique, and S was calculated using a plot of the thermoelectric potential as a function of the temperature difference.

3. Results and Discussion

3.1 Phase identification

The XRD patterns of the powders prepared via ball milling are shown in Fig. 1. Compared to the XRD pattern of the arc-melted Fe₂V_1.08Al_0.92 alloy before ball milling (Fig. 1(a)), diffraction peaks representing the L2₁ structure, e.g., at approximately 2θ = 26.7°, corresponding to 111, and B2 ordering, e.g., at 2θ = 31.0°, corresponding to 200, are no longer observed, with only the peaks representing body-centered cubic (bcc) structures remaining in the XRD patterns of the ball-milled powders. Broadening of the diffraction peaks is also observed. These results suggest that a sufficient mechanical grinding effect is applied during ball milling to reduce the crystallite size and even disturb the crystal structure. Notably, the particles in the powder prepared via ball milling comprise nanometer-sized agglomerated flakes, as observed when preparing Fe₂VAl powder using mechanical alloying [13]. The broadening of the peaks representing the Fe₂VAl phase is slightly enhanced by the addition of the oxide materials. The estimated crystallite sizes of the Fe₂VAl phase within the powders with added oxides, based on the Scherrer equation, are reduced from 13 (FVA) to 11 nm. The diffraction peaks due to the Al₂O₃ or La₂O₃ additives are not observed, mainly because their diffraction intensities are too weak due to their low contents.

Fig. 1

XRD patterns of the (a) Fe₂V_1.08Al_0.92 alloy prepared via arc melting and the (b) FVA, (c) mAO, (d) nAO, (e) mLO, (f) nLO-10, and (g) nLO-20 powders.

The XRD patterns of the sintered samples are shown in Fig. 2. The diffraction peaks representing L2₁ ordering are observed in the patterns of all samples, confirming that the crystal structure of the ball-milled powder changed from disordered to L2₁ ordered one during the sintering. Additional peaks representing secondary or oxide phases are not observed. As shown in Table 1, the lattice parameter of sintered FVA, as calculated using the peak positions in the XRD pattern, is consistent with that of the arc-melted sample. In addition, the lattice parameters of the sintered samples of mAO and nAO are similar to the original value of the arc-melted sample. As the lattice parameter of the Heusler Fe₂VAl phase is sensitive to the degree of L2₁ ordering [23], these results suggest that the Heusler Fe₂VAl phase disturbed by ball milling is almost completely recovered. However, the lattice parameters of nLO-10 and nLO-20 generally increase, suggesting that La₂O₃ reacts with the Fe₂VAl phase. The solid solution of La, with a larger covalent radius than those of the constituents of the Fe₂VAl alloy, is likely to increase the lattice parameter of the alloy. However, further investigation is necessary to clearly reveal the cause of the increased lattice parameter.

Fig. 2

XRD patterns of the (a) Fe₂V_1.08Al_0.92 alloy prepared via arc melting and the sintered (b) FVA, (c) mAO, (d) nAO, (e) mLO, (f) nLO-10, and (g) nLO-20 powders.

Table 1 Lattice parameters of the sintered Fe₂VAl alloys.

3.2 Characterization of microstructure formation

The SEM images of the fractured surfaces of the sintered samples and the evaluated grain sizes based on the images are shown in Figs. 3 and 4, respectively. Owing to the combination of the reduction in crystallite size due to ball milling and rapid PECS, FVA exhibits a submicrometer-sized microstructure, as shown in Fig. 3(a). The matrix grain size of mAO is almost of the same order as that of FVA. However, the dispersion of particles with sizes of several tens of nanometers, which are apparently formed via the milling of the Al₂O₃ particles, is observed. This suggests that the sparsely distributed coarse Al₂O₃ particles display little influence on grain growth during sintering. The grain size of nAO, however, is apparently reduced, which may be simply explained by the pinning effects of the well-dispersed nanometer-sized Al₂O₃ particles that inhibit grain growth during sintering. In contrast, when La₂O₃ is added, several sections of the sintered body comprising considerably smaller grains may be observed, regardless of the particle size of the added La₂O₃ powder, as shown in Figs. 3(d) and 3(e). This difference in the effects of Al₂O₃ and La₂O₃ addition on the microstructure may be attributed to the lower stability of La₂O₃, and the friable La₂O₃ may be easily dispersed via the mechanical grinding effect of ball milling. In addition, the slight increases in the lattice parameters of the Fe₂VAl phases in the La₂O₃-doped samples indicate that La₂O₃ partially reacts with the Fe₂VAl phase, which may enhance the dispersion of La₂O₃ into the Fe₂VAl phase. The finely dispersed La₂O₃ particles and secondary phase derived from the reaction between Fe₂VAl and La₂O₃ may reduce grain growth during sintering via the drag effect upon grain boundary segregation. In addition, regions comprising relatively coarse grains with sizes of 1–2 µm are observed in mLO and nLO-10. As this grain growth to the micrometer size is caused by a lack of La₂O₃ dispersion, the relatively coarse grains may be refined by increasing the duration of ball milling to enhance La₂O₃ dispersion. Therefore, in nLO-20, almost all regions comprise nanometer-sized grains, as shown in Fig. 3(f).

Fig. 3

SEM images of the fracture cross-sections of the sintered (a) FVA, (b) mAO, (c) nAO, (d) mLO, (e) nLO-10, and (f) nLO-20 powders.

Fig. 4

Evaluated grain sizes of the sintered samples, based on the SEM images.

3.3 Evaluation of the thermoelectric properties

To assess the influences of microstructure refinement and oxide particle addition on the thermoelectric performance, the thermoelectric properties were evaluated. As shown in Fig. 5, FVA exhibits a larger |S| than that of the arc-melted sample. In a previous study, the Fe₂V_1.08Al_0.92 alloy with slight hole doping exhibited a similar behavior, with an increase in |S| accompanied by a decrease in the peak temperature [2]. Therefore, the electron-doping effect of off-stoichiometric compositional control of the Fe₂V_1.08Al_0.92 alloy is apparently partially canceled out by the hole-doping effect. As the S values of Fe₂VAl alloys are sensitive to small compositional changes [1–3], unavoidable oxidation and a slight compositional variation in the Fe₂VAl phase during ball milling should cause the increased |S| of FVA. mAO and nAO also exhibit the same trend as that of FVA, and thus, Al₂O₃ addition displays little influence on the electronic state of the Fe₂VAl alloy. Conversely, the samples with added La₂O₃ exhibit lower |S| values compared to that of FVA. As La₂O₃ is likely to partially react with the Fe₂V_1.08Al_0.92 alloy, La doping of the Fe₂V_1.08Al_0.92 phase should affect its electronic state. In previous studies, the decrease in |S| due to electron doping of the Fe₂VAl system is typically accompanied by an increase in the peak temperature of |S| [1]. Therefore, the decrease in |S| over the entire temperature range of measurement (without a distinct increase in the peak temperature) due to La₂O₃ addition is unlikely to be caused by the electron-doping effect. The mechanism of the decrease in |S| via La₂O₃ addition remains unknown. Nevertheless, the reaction with La₂O₃ apparently induces the metallic nature of the Fe₂V_1.08Al_0.92 phase due to the distortion of the pseudogap electronic structure of the Fe₂VAl phase.

Fig. 5

Temperature dependences of the Seebeck coefficients.

The temperature dependences of the σ values of the Fe₂VAl samples are shown in Fig. 6. Compared to that of the arc-melted Fe₂V_1.08Al_0.92 alloy, every sintered Fe₂VAl sample exhibits a lower σ, mainly due to the increased number of grain boundaries, which scatters electron transport. mAO and nAO exhibit almost identical orders of σ compared to that of FVA, although the dispersion of particles with sizes of several tens of nanometers (in mAO) and the reduction of the grain size (in nAO) are observed. As the reported length of the electron mean free path in the Fe₂VAl alloy is 2.4 nm [24], the reduction of the grain size in the submicrometer range displays little influence on electron transport. The σ values of nLO-10 and nLO-20 are slightly reduced, particularly at low temperatures. The lattice distortion, which is represented by a scattering factor in the nanometer range, induced by the reaction of La₂O₃ with the Fe₂VAl phase scatters electron transport, resulting in the lower σ. The σ of mLO is enhanced, particularly at low temperatures, and the combined effect of the remaining coarse grains and induced metallic nature due to the reaction with La₂O₃ contributes in enhancing σ.

Fig. 6

Temperature dependences of the electrical conductivities.

These S and σ values suggest that Al₂O₃ addition exhibits little effect on the electronic component of the thermoelectric properties of the Fe₂VAl alloy. The calculated thermoelectric PF values (PF = S²σ) of mAO and nAO are at the same level as that of FVA, as shown in Fig. 7. Conversely, La₂O₃ addition negatively affects the electronic component of thermoelectric performance, mainly due to the decrease in |S|. The negative influence increases with an increase in the dispersion of La₂O₃, and this trend is clearly observed in the PF values, as shown in Fig. 7. The PF of nLO-20 is less than half of that of FVA, and thus, in terms of the electronic component of thermoelectric performance, La₂O₃ addition is not favorable.

Fig. 7

Temperature dependences of the power factors.

The temperature dependences of the κ values and lattice components of the κ values (κ_ph) of the Fe₂VAl samples are shown in Figs. 8 and 9. κ_ph is calculated by subtracting the electronic component of κ, which is estimated using the Wiedemann-Franz law with a Lorenz number of 2.44 × 10⁻⁸ WΩ/K², from the total κ. The reported cumulative κ_ph of the Fe₂VAl alloy suggests that most of the lengths of the phonon mean free paths are in the range 20–300 nm [18]. Therefore, the κ_ph values of nAO and the La₂O₃-doped samples, which comprise grains with sizes of nanometers, are strongly reduced compared to that of FVA with grain sizes of submicrometers. Remarkably, mLO and nLO-20 exhibit almost identical levels of κ_ph, although their microstructures are rather different. The uniformly reduced grain size over the entire sintered body of nLO-20 is the sole cause of its reduced phonon transport. In contrast, the component of the sintered body of mLO comprising fine grains may be mainly responsible for its reduced phonon transport, whereas the rest of the sintered body comprising coarse grains displays a lower influence on phonon transport, as for FVA. As La₂O₃ addition inherently negatively affects the electronic component of the thermoelectric property, the latter case favors the enhancement of the thermoelectric performance of the Fe₂VAl alloy.

Fig. 8

Temperature dependences of the thermal conductivities.

Fig. 9

Temperature dependences of the lattice components of the thermal conductivities, as estimated using the Wiedemann-Franz law.

3.4 Assessment of thermoelectric performance and power generation capacity

To evaluate the thermoelectric performances of the sintered samples, their ZT values were calculated, as shown in Fig. 10. Compared to that of arc-melted Fe₂V_1.08Al_0.92, the ZT of FVA is decreased due to the negative effects of the grain boundaries on electron transport. The ZT of nAO, however, is comparable to that of arc-melted Fe₂V_1.08Al_0.92 because the negative effects of the grain boundaries on electron transport are compensated by phonon scattering due to microstructure refinement. In addition, mLO exhibits a slightly higher ZT than that of arc-melted Fe₂V_1.08Al_0.92, despite La₂O₃ addition inherently negatively affecting the thermoelectric performance, as indicated by the low ZT of nLO-20. This suggests that partial microstructure refinement may improve the thermoelectric performance by controlling the balance of the effects on electronic and phonon transport. Therefore, further investigations of detailed control over the microstructure, including the volume ratio and distribution of the fine-grained component, may yield an Fe₂VAl alloy with an enhanced thermoelectric performance.

Fig. 10

Temperature dependences of the thermoelectric figures of merit.

The effects of the modifications of the thermoelectric properties caused by microstructure refinement due to oxide particle addition on thermoelectric power generation were assessed. To this end, the power densities and conversion efficiencies of arc-melted Fe₂V_1.08Al_0.92 and mLO were estimated using their measured thermoelectric properties, as shown in Fig. 11. A simple model with a fixed-temperature heat source and sink at 773 K and 293 K, respectively, and an assumed thermal resistance of 10 K/W between the heat sources and thermoelectric leg was used in these estimations. By changing the leg length, the output power density may be maximized by adjusting the balance between the thermal and electrical resistances of the leg. The thermal resistance should be large to yield a large thermoelectric voltage due to the temperature difference, whereas the electrical resistance should be reduced to decrease the power loss due to current flow through the leg. As shown in Fig. 11, the output power density of mLO is higher than that of arc-melted Fe₂V_1.08Al_0.92, although arc-melted Fe₂V_1.08Al_0.92 displays a higher PF. Hence, a lower κ can contribute in enhancing the electronic characteristics of a thermoelectric device via the enlargement of the temperature difference at the leg due to an increased thermal resistance. In addition, a lower κ can contribute in enhancing the conversion efficiency of mLO because of the reduced heat flow through the leg. Moreover, the lower κ of mLO can further contribute in reducing the optimal leg length to maximize the power density, as shown in Fig. 11. These results suggest that a reduction in κ via microstructure refinement is favorable in enhancing the power generation capacity and reducing the size and mass of a thermoelectric power generation device composed of the Fe₂VAl alloy. Therefore, from a practical perspective, a reduction in κ should contribute in reducing the cost and enhancing the adaptability of a thermoelectric power generation system to fit the limited spaces around heat sources.

Fig. 11

Estimated power densities and conversion efficiencies of thermoelectric power generation based on the thermoelectric properties of the Fe₂V_1.08Al_0.92 alloy prepared via arc melting and sintered mLO. Estimation was performed using a fixed-temperature heat source and sink at 773 K and 293 K, respectively, assuming 10 K/W of thermal resistance between the thermoelectric leg and heat sources.

4. Conclusion

The microstructure of the sintered Fe₂V_1.08Al_0.92 alloy was modified by adding oxide particles. By dispersing the oxide nanoparticles, the microstructure was refined to a nanometer size. Upon addition of the Al₂O₃ powder with a mean particle size of <13 nm, the grain size of the Fe₂V_1.08Al_0.92 alloy was reduced to approximately 200 nm and the κ_ph of arc-melted Fe₂V_1.08Al_0.92 of 11.7 W/mK was suppressed to 8.7 W/mK. This reduction in κ_ph could compensate for the negative effects of microstructure refinement on the electronic properties due to electron scattering at the grain boundaries, and the ZT of arc-melted Fe₂V_1.08Al_0.92 was thus almost reproduced. Upon addition of the La₂O₃ powder, a sintered Fe₂V_1.08Al_0.92 alloy with fine grains of approximately 100 nm was obtained, regardless of the particle size of the added La₂O₃ powder, due to the reaction between Fe₂V_1.08Al_0.92 and La₂O₃. As La₂O₃ dispersion into the Fe₂VAl alloy inherently negatively influenced the electronic component of the thermoelectric property of the Fe₂VAl alloy, mainly due to the decrease in |S|, improving the ZT of the Fe₂VAl alloy via La₂O₃ addition was challenging. However, the appreciable reduction in κ_ph without a considerable degradation in the electronic properties due to partial microstructure refinement exhibited the potential to improve the thermoelectric performance of the Fe₂VAl alloy. Upon addition of the La₂O₃ powder with a mean particle size of 5 µm, the ZT of Fe₂V_1.08Al_0.92 was slightly improved.

These microstructural modifications could not significantly improve the thermoelectric performance of the Fe₂VAl alloy. However, the reduction in κ without deteriorating ZT could enhance the output power density and reduce the optimal leg length of the thermoelectric module for use in thermoelectric power generation. The power density was estimated using a simple model with a fixed-temperature heat source and heat sink at 773 and 293 K, respectively, assuming a thermal resistance of 10 K/W between the heat sources and thermoelectric leg. The results suggested that the output power density of 0.62 W/cm² of the arc-melted Fe₂V_1.08Al_0.92 alloy could be enhanced to 0.68 W/cm² using the sintered La₂O₃-doped Fe₂V_1.08Al_0.92 alloy, although arc-melted Fe₂V_1.08Al_0.92 displayed a higher PF. In addition, the optimal leg length of 7 mm of the arc-melted Fe₂V_1.08Al_0.92 alloy could be reduced to 5 mm using the sintered La₂O₃-doped Fe₂V_1.08Al_0.92 alloy. Therefore, controlling the balance between the electronic and thermal properties via microstructural modification is favorable in improving the practicability of the Fe₂VAl alloy by enhancing the power generation capacity and reducing the sizes and masses of thermoelectric devices.

These conclusions may be summarized as follows:

1. (1)    The grain growth during the sintering of the Heusler Fe₂V_1.08Al_0.92 alloy was suppressed by the dispersion of oxide nanoparticles.
2. (2)    The nanometer-sized refined microstructure reduced the κ_ph and σ of the Heusler Fe₂V_1.08Al_0.92 alloy without deteriorating ZT.
3. (3)    The suppression of the high κ of the Fe₂VAl alloy as a thermoelectric material could reduce the size of the thermoelectric module and enhance the output power density in thermoelectric power generation.
4. (4)    Controlling the balance between the electronic and thermal properties via microstructural modification could facilitate the practical application of thermoelectric materials.

Acknowledgments

This work was partially supported by the Adaptable and Seamless Technology Transfer Program through Target-driven R&D, A-STEP (Nos. AS2415009L and AS2916001), and the Japan Science and Technology Agency.

REFERENCES

• 1)   H.  Miyazaki,  S.  Tanaka,  N.  Ide,  K.  Soda and  Y.  Nishino: Thermoelectric properties of Heusler-type off-stoichiometric Fe₂V_1+xAl_1−x alloys, Mater. Res. Express 1 (2014) 015901. doi:10.1088/2053-1591/1/1/015901
• 2)   Y.  Nishino,  S.  Kamizono,  H.  Miyazaki and  K.  Kimura: Effects of off-stoichiometry and Ti doping on thermoelectric performance of Fe₂VAl Heusler compound, AIP Adv. 9 (2019) 125003. doi:10.1063/1.5123783
• 3)   F.  Garmroudi,  A.  Riss,  M.  Parzer,  N.  Reumann,  H.  Müller,  E.  Bauer,  S.  Khmelevskyi,  R.  Podloucky,  T.  Mori,  K.  Tobita,  Y.  Katsura and  K.  Kimura: Boosting the thermoelectric performance of Fe₂VAl-type Heusler compounds by band engineering, Phys. Rev. B 103 (2021) 085202. doi:10.1103/PhysRevB.103.085202
• 4)   M.  Mikami,  K.  Kobayashi,  T.  Kawada,  K.  Kubo and  N.  Uchiyama: Development and Evaluation of High-Strength Fe₂VAl Thermoelectric Module, Jpn. J. Appl. Phys. 47 (2008) 1512–1516. doi:10.1143/JJAP.47.1512
• 5)   K.  Abe,  A.  Kikuchi,  N.  Okinaka and  T.  Akiyama: Single thermite-type combustion synthesis of Fe₂VAl for thermoelectric applications from Fe, V₂O₅ and Al powders, J. Alloy. Compd. 611 (2014) 319–323. doi:10.1016/j.jallcom.2014.05.137
• 6)   C.  van der Rest,  V.  Dupont,  J.-P.  Erauw and  P.J.  Jacques: On the reactive sintering of Heusler Fe₂VAl-based thermoelectric compounds, Intermetallics 125 (2020) 106890. doi:10.1016/j.intermet.2020.106890
• 7)   M.  Mikami,  Y.  Kinemuchi,  K.  Kubo,  N.  Uchiyama,  H.  Miyazaki and  Y.  Nishino: Near-Net-Shape Fabrication of Thermoelectric Legs by Flash Sintering, J. Electron. Mater. 49 (2020) 593–600. doi:10.1007/s11664-019-07743-0
• 8)   Y.  Nishino,  S.  Deguchi and  U.  Mizutani: Thermal and transport properties of the Heusler-type Fe₂VAl_1−x Ge_x (0 ≤ x ≤ 0.20) alloys: Effect of doping on lattice thermal conductivity, electrical resistivity, and Seebeck coefficient, Phys. Rev. B 74 (2006) 115115. doi:10.1103/PhysRevB.74.115115
• 9)   M.  Mikami,  Y.  Kinemuchi,  K.  Ozaki,  Y.  Terazawa and  T.  Takeuchi: Thermoelectric properties of tungsten-substituted Heusler Fe₂VAl alloy, J. Appl. Phys. 111 (2012) 093710. doi:10.1063/1.4710990
• 10)   T.  Takeuchi,  Y.  Terazawa,  Y.  Furuta,  A.  Yamamoto and  M.  Mikami: Effect of Heavy Element Substitution and Off-Stoichiometric Composition on Thermoelectric Properties of Fe₂VAl-Based Heusler Phase, J. Electron. Mater. 42 (2013) 2084–2090. doi:10.1007/s11664-013-2532-0
• 11)   K.  Renard,  A.  Mori,  Y.  Yamada,  S.  Tanaka,  H.  Miyazaki and  Y.  Nishino: Thermoelectric properties of the Heusler-type Fe₂VTa_xAl_1−x alloys, J. Appl. Phys. 115 (2014) 033707. doi:10.1063/1.4861419
• 12)   A.  Diack-Rasselio,  O.  Rouleau,  L.  Coulomb,  L.  Georgeton,  M.  Beaudhuin,  J.-C.  Crivello and  E.  Alleno: Influence of self-substitution on the thermoelectric Fe₂VAl Heusler alloy, J. Alloy. Compd. 920 (2022) 166037. doi:10.1016/j.jallcom.2022.166037
• 13)   M.  Mikami,  A.  Matsumoto and  K.  Kobayashi: Synthesis and thermoelectric properties of microstructural Heusler Fe₂VAl alloy, J. Alloy. Compd. 461 (2008) 423–426. doi:10.1016/j.jallcom.2007.07.004
• 14)   M.  Mikami and  K.  Kobayashi: Effect of Bi addition on microstructure and thermoelectric properties of Heusler Fe₂VAl-sintered alloy, J. Alloy. Compd. 466 (2008) 530–534. doi:10.1016/j.jallcom.2007.11.117
• 15)   S.  Masuda,  K.  Tsuchiya,  J.  Qiang,  H.  Miyazaki and  Y.  Nishino: Effect of high-pressure torsion on the microstructure and thermoelectric properties of Fe₂VAl-based compounds, J. Appl. Phys. 124 (2018) 035106. doi:10.1063/1.5034390
• 16)   L.  Gomell,  S.  Katnagallu,  A.  Diack-Rasselio,  S.  Maier,  L.  Perrière,  C.  Scheu,  E.  Alleno and  B.  Gault: Chemical segregation and precipitation at anti-phase boundaries in thermoelectric Heusler-Fe₂VAl, Scr. Mater. 186 (2020) 370–374. doi:10.1016/j.scriptamat.2020.04.037
• 17)   L.  Gomell,  T.  Haeger,  M.  Roscher,  H.  Bishara,  R.  Heiderhoff,  T.  Riedl,  C.  Scheu and  B.  Gault: Microstructure manipulation by laser-surface remelting of a full-Heusler compound to enhance thermoelectric properties, Acta Mater. 223 (2022) 117501. doi:10.1016/j.actamat.2021.117501
• 18)   N.  Fukatani,  Y.  Kurosaki,  S.  Yabuuchi,  A.  Nishide and  J.  Hayakawa: Improved power factor in low thermal conductive Fe₂VAl-based full-Heusler thin films by composition-control with off-axis sputtering method, Appl. Phys. Lett. 112 (2018) 033902. doi:10.1063/1.5012106
• 19)   P.A.  Manohar,  M.  Ferry and  T.  Chandra: Five Decades of the Zener Equation, ISIJ Int. 38 (1998) 913–924. doi:10.2355/isijinternational.38.913
• 20)   M.  Hillert: Inhibition of grain growth by second-phase particles, Acta Metall. 36 (1988) 3177–3181. doi:10.1016/0001-6160(88)90053-3
• 21)   A.  Harun,  E.A.  Holm,  M.P.  Clode and  M.A.  Miodownik: On computer simulation methods to model Zener pinning, Acta Mater. 54 (2006) 3261–3273. doi:10.1016/j.actamat.2006.03.012
• 22)   N.  Moelans,  B.  Blanpain and  P.  Wollants: Pinning effect of second-phase particles on grain growth in polycrystalline films studied by 3-D phase field simulations, Acta Mater. 55 (2007) 2173–2182. doi:10.1016/j.actamat.2006.11.018
• 23)   Y.  Nishino: Unusual electron transport in Heusler-type Fe₂VAl compound, Intermetallics 8 (2000) 1233–1241. doi:10.1016/S0966-9795(00)00037-6
• 24)   Y.  Feng,  J.Y.  Rhee,  T.A.  Wiener,  D.W.  Lynch,  B.E.  Hubbard,  A.J.  Sievers,  D.L.  Schlagel,  T.A.  Lograsso and  L.L.  Miller: Physical properties of Heusler-like Fe₂VAl, Phys. Rev. B 63 (2001) 165109. doi:10.1103/PhysRevB.63.165109