Thermoelectric conversion device

Abstract

To provide a thermoelectric conversion element and a thermoelectric conversion device made of inexpensive and non-toxic materials.SOLUTION: A thermoelectric conversion element is made of: a first material (e.g., Fe3Al) with a stoichiometric composition in which a composition formula is expressed by Fe3X and X is a typical element or a transition element; a second material with an off-stoichiometric composition in which a composition ratio of Fe to X deviates from that of the first material; a third material obtained by substituting a part of Fe sites in the first material or a part of Fe sites in the second material by a typical metal element or a transition element other than X; a fourth material in which a composition formula is expressed by Fe3M11-xM2x(0<x<1), and M1 and M2 are different from each other; or a fifth material obtained by substituting a part of Fe sites in the first material by a transition element other than X and substituting a part of X sites by a typical metal element other than X. The first material, the second material, the third material, the fourth material and the fifth material exhibit an anomalous Nernst effect.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a thermoelectric conversion element and a thermoelectric conversion device including the thermoelectric conversion element. [Background technology]

[0002] The Seebeck effect is known as a thermoelectric mechanism that generates a voltage when a temperature gradient is applied to a substance (see, for example, Patent Document 1). However, in a thermoelectric mechanism using the Seebeck effect, the materials that can be used above room temperature are mainly made of bismuth, tellurium, lead, etc., which are highly toxic and therefore not suitable for practical use, and are mechanically fragile, vulnerable to vibration, and not durable. In addition, since the Seebeck effect generates a voltage in the same direction as the temperature gradient, it is necessary to fabricate a three-dimensional, complex structure in which p-type modules and n-type modules are alternately arranged in the vertical direction from the heat source surface, which results in high manufacturing costs. In addition, it is difficult to deploy such a three-dimensional element over a large area.

[0003] Similarly, the anomalous Nernst effect is known as a thermoelectric mechanism that generates a voltage due to a temperature gradient. The anomalous Nernst effect is a phenomenon in which, when a heat flow is passed through a magnetic material, causing a temperature difference, a voltage is generated in a direction perpendicular to both the magnetization direction and the temperature gradient. In recent years, it has been discovered that by utilizing the topology of the electronic structure, the Nernst coefficient can be increased much more than the previously known value (0.1 μV / K). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2016 / 181777 Summary of the Invention [Problem to be solved by the invention]

[0005] However, although various magnetic materials showing an anomalous Nernst effect have been developed so far, there has been a problem of high cost because relatively expensive metals are used.

[0006] The present invention has been made in view of the above problems, and an object thereof is to provide a thermoelectric conversion element and a thermoelectric conversion device made of an inexpensive and non-toxic material.

Means for Solving the Problems

[0007] The thermoelectric conversion element according to an embodiment of the present invention includes a first substance having a stoichiometric composition represented by the composition formula Fe3X, where X is a typical element or a transition element, a second substance having an off-stoichiometric composition in which the composition ratio of Fe and the X deviates from that of the first substance, a third substance in which a part of the Fe sites of the first substance or a part of the Fe sites of the second substance is substituted with a typical metal element or a transition element other than the X, a fourth substance represented by the composition formula Fe3M1 1-x M2 x (0 < x < 1), where M1 and M2 are different typical elements from each other, or a fifth substance in which a part of the Fe sites of the first substance is substituted with a transition element other than X and a part of the X sites is substituted with a typical metal element other than X. The first substance, the second substance, the third substance, the fourth substance, and the fifth substance exhibit an anomalous Nernst effect.

[0008] The thermoelectric conversion device according to an embodiment of the present invention includes a substrate and a power generator provided on the substrate and having a plurality of thermoelectric conversion elements. Each of the plurality of thermoelectric conversion elements has a shape extending in one direction and is made of the above-described first substance, second substance, third substance, fourth substance, or fifth substance. The plurality of thermoelectric conversion elements are arranged in parallel in a direction perpendicular to the one direction and are electrically connected in series.

[0009] The thermoelectric conversion device according to another embodiment of the present invention includes the above-described thermoelectric conversion element and a hollow member. The thermoelectric conversion element is in a sheet-like structure or a wire and is provided so as to cover the outer surface of the hollow member.

Effects of the Invention

[0010] According to the present invention, the anomalous Nernst effect can be realized by a thermoelectric conversion element made of inexpensive and non-toxic materials. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a schematic diagram showing the crystal structure of D03-type Fe3X and the crystal structure of L12-type Fe3X. [Diagram 2] 1 is a schematic diagram for explaining a thermoelectric mechanism of a thermoelectric conversion element according to the present embodiment. FIG. [Diagram 3] 1 is a graph showing the temperature dependence of the Nernst coefficient in various metal materials. [Figure 4] 4 is a graph in which the Nernst coefficients shown in FIG. 3 are normalized by the Nernst coefficient at T=300K. [Diagram 5] 1 is a graph showing the magnetic field dependence of the Nernst coefficient at T=300K of a single crystal and polycrystalline Fe—Al alloy, and a polycrystalline Fe—Al—V alloy. [Figure 6] 1 is a graph showing the temperature dependence of the Nernst coefficient of a single crystal and polycrystalline Fe—Al alloy, and a polycrystalline Fe—Al—V alloy. [Figure 7] 1 is a perspective view showing a configuration of a thermoelectric conversion device according to Example 1 including a thermoelectric conversion element of the present embodiment. [Figure 8] FIG. 11 is a plan view showing a configuration of a thermoelectric conversion device according to Example 2 including the thermoelectric conversion element of the present embodiment. [Figure 9] FIG. 11 is an external view showing a configuration of a thermoelectric conversion device according to Example 3 including the thermoelectric conversion element of the present embodiment. [Figure 10] 1 is a graph showing the magnetic field dependence of the Nernst coefficients at T=300K for Fe3Pt, Fe3Ge, Fe3Al, Fe3Ga0.5Al0.5, Fe3Ga, and Co2MnGa. [Figure 11] 1 is a graph showing the temperature dependence of the Nernst coefficients of Fe3Pt, Fe3Ge, Fe3Al, Fe3Ga, and Co2MnGa. [Figure 12A] 1 is a graph showing the temperature dependence of the Nernst coefficient of Fe3Sn. [Figure 12B] 1 is a graph showing the temperature dependence of the transverse thermoelectric conductivity of Fe3Sn. [Figure 13A] 1 is a table showing the Nernst coefficients at T=300K for the mixed crystal system of Fe3Si and Fe3Al. [Figure 13B] 1 is a graph showing the Nernst coefficient at T=300K for a mixed crystal system of Fe3Si and Fe3Al. [Figure 14A] 1 is a table showing the Nernst coefficients at T=300K for a mixed crystal system of Fe3Al and Fe3Ga. [Figure 14B] 1 is a graph showing the Nernst coefficient at T=300K for a mixed crystal system of Fe3Al and Fe3Ga. [Figure 15] 1 is a graph showing the Nernst coefficient at T=300K of Fe3Cu1-xGax single crystal and polycrystal obtained by substituting a part of the Ga site of Fe3Ga with Cu. [Figure 16A] 1 is a graph showing X-ray diffraction patterns of Nd0.1Fe2.9Ga and Fe3Ga. [Figure 16B] 1 is a graph showing the magnetic field dependence of magnetization of Nd0.1Fe2.9Ga at T=300K. [Figure 16C] 1 is a graph showing the magnetic field dependence of the Nernst coefficient at T=300K for an Fe3Ga single crystal and an Nd0.1Fe2.9Ga polycrystal. [Figure 16D] 1 is a graph showing the magnetic field dependence of Hall resistivity at T=300K of an Fe3Ga single crystal and an Nd0.1Fe2.9Ga polycrystal. [Figure 17A] 1 is a graph showing X-ray diffraction patterns of Ho0.05Fe2.95Ga and Fe3Ga. [Figure 17B] 1 is a graph showing the magnetic field dependence of magnetization of Ho0.05Fe2.95Ga at T=300K. [Figure 18A] 1 is a graph showing X-ray diffraction patterns of Y0.05Fe2.95Ga and Fe3Ga. [Figure 18B] 1 is a graph showing the magnetic field dependence of magnetization of Y0.05Fe2.95Ga. [Figure 19A]1 is a graph showing the magnetic field dependence of magnetization of Tb0.05Fe2.95Ga. [Figure 19B] 19B is a graph showing an enlarged view of the low magnetic field area of ​​FIG. 19A. [Figure 20A] 1 is a graph showing X-ray diffraction patterns of Tb0.03Fe2.97Ga and Fe3Ga. [Figure 20B] 1 is a graph showing the magnetic field dependence of magnetization of Tb0.03Fe2.97Ga at T=300K. [Figure 21A] 1 is a graph showing X-ray diffraction patterns of Fe3Ga0.8B0.2, Fe3Ga0.9B0.1, and Fe3Ga. [Figure 21B] 1 is a graph showing the magnetic field dependence of magnetization at T=300 K for a needle-shaped sample and a plate-shaped sample made of Fe3Ga0.8B0.2. [Figure 21C] 1 is a graph showing the magnetic field dependence of magnetization of needle-shaped samples of Fe3Ga0.8B0.2 and Fe3Ga. [Figure 21D] 1 is a graph showing the magnetic field dependence of magnetization of Fe3Ga0.8B0.2 and Fe3Ga of plate-shaped samples. [Figure 21E] 1 is a graph showing the magnetic field dependence of the Nernst coefficient of a plate-shaped sample of Fe3Ga0.8B0.2. [Figure 21F] 1 is a graph showing the magnetic field dependence of Hall resistivity of a plate-shaped sample of Fe3Ga0.8B0.2. [Figure 22A] 1 is a graph showing X-ray diffraction patterns of Fe2.9Mn0.1Ga, Fe2.5Mn0.5Ga, Fe2MnGa, and Fe3Ga. [Figure 22B] 1 is a graph showing the magnetic field dependence of magnetization at T=300 K for needle-shaped samples of Fe2.9Mn0.1Ga and Fe2.5Mn0.5Ga. [Figure 23A] 1 is a graph showing X-ray diffraction patterns of Fe2.9Pt0.1Ga, Fe2.9Pt0.1Ga0.9Ge0.1, and Fe3Ga. [Figure 23B] 1 is a graph showing the magnetic field dependence of magnetization at T=300 K for needle-shaped samples of Fe2.9Pt0.1Ga and Fe2.9Pt0.1Ga0.9Ge0.1. [Figure 24] FIG. 1 is a schematic diagram for explaining the anomalous Nernst effect when a temperature gradient is applied in the in-plane direction of a thin film sample as a thermoelectric conversion element. [Diagram 25] 1 is a graph showing the measurement results of the anomalous Nernst effect at T=300K when a temperature gradient is applied in the in-plane direction to a thin film sample (Fe3Ga). [Figure 26] FIG. 1 is a schematic diagram for explaining the anomalous Nernst effect when a temperature gradient is applied in a direction perpendicular to the surface of a thin film sample as a thermoelectric conversion element. [Figure 27A] FIG. 1 is a schematic diagram for explaining a method for measuring the anomalous Nernst effect when a temperature gradient is applied to a thin film sample in a direction perpendicular to the surface. [Figure 27B] FIG. 1 is a schematic diagram for explaining a method for measuring the anomalous Nernst effect when a temperature gradient is applied to a thin film sample in a direction perpendicular to the surface. [Figure 28] 1 is a graph showing the measurement results of the anomalous Nernst effect when a temperature gradient is applied in the direction perpendicular to the surface of a thin film sample (Fe3Ga). [Figure 29A] 1 is a graph showing the measurement results of the anomalous Nernst effect when a temperature gradient is applied in the direction perpendicular to the surface of a thin film sample of Fe3Ga obtained without annealing after film formation at room temperature. [Figure 29B] 29B is a graph showing an enlarged view of the low magnetic field area of ​​FIG. 29A. [Diagram 30] 1 is a graph showing the measurement results of the anomalous Nernst effect when a temperature gradient is applied in the direction perpendicular to the surface for an Fe3Ga epitaxial film obtained by annealing after film formation at room temperature, and an Fe3Ga amorphous film obtained by not annealing after film formation at room temperature. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Exemplary embodiments of the present invention will now be described with reference to the accompanying drawings.

[0013] Among materials that exhibit the anomalous Nernst effect, the highest Nernst coefficient at room temperature to date is 6 μV / K, which the inventors of the present application achieved in Co2MnGa (see Nature Physics 14, 1119-1124 (2018) and WO 2019 / 009308).

[0014] As described below, the inventors of the present application have been able to achieve a Nernst coefficient approaching the highest value ever achieved with the binary Fe3Al system. The Clarke number, which indicates the weight ratio of elements present near the Earth's surface, is known to be oxygen (O), silicon (Si), aluminum (Al), iron (Fe), and so on, in descending order. As such, Fe and Al have relatively large Clarke numbers, and are therefore very inexpensive materials that are also non-toxic. Fe3Al is also chemically stable, and has a high ferromagnetic transition temperature of approximately 700K.

[0015] The crystal structure of Fe3X (X is a typical or transition element) that can exhibit the anomalous Nernst effect is shown in Figure 1. As shown in Figure 1, Fe3X can have a D03 type structure (a) and an L12 type structure (b).

[0016] The unit cell of the D03 structure (a) has eight body-centered cubic (bcc) type subcells. In each subcell, Fe atoms (Fe(II)) occupy the corners, and each Fe(II) is shared by the eight adjacent subcells. Four of the eight subcells have four Fe atoms (Fe(I)) occupying the body centers, and the remaining four subcells have four X atoms occupying the body centers. For example, the lattice constant a of Fe3Al in the D03 structure is 5.64 Å (see Physical Review B 66, 205203 (2002)).

[0017] The L12 structure (b) is a face-centered cubic (fcc) type crystal structure in which Fe atoms are located at the face centers and X atoms are located at the corners.

[0018] For example, Fe3Al single crystals were produced by arc-melting Fe and Al in the appropriate ratio, growing the crystal by pulling it using the Czochralski method, annealing the crystal at a low temperature (e.g., 500°C), and slowly cooling it to room temperature over a few minutes to a few tens of minutes. Electron diffraction revealed that the produced Fe3Al single crystal was an ordered phase (D03 phase (Fm-3m)).

[0019] On the other hand, Fe3Al polycrystals were prepared by arc-melting Fe and Al in an appropriate ratio to prepare a polycrystalline sample, annealing the sample at a high temperature (e.g., 900°C), and quenching it to room temperature in a few seconds. From the phase diagram, the prepared Fe3Al polycrystal is considered to be a disordered phase (B2 phase (Pm-3m) or A2 phase (Im-3m)) or a mixed crystal with an ordered phase.

[0020] Next, a thermoelectric conversion element and its thermoelectric mechanism according to an embodiment of the present invention will be described with reference to FIG.

[0021] The thermoelectric conversion element 1 according to this embodiment is made of Fe3X single crystal or polycrystal produced by the above-mentioned method. As shown in FIG. 2, the thermoelectric conversion element 1 is a rectangular parallelepiped extending in one direction (y direction), has a predetermined thickness (length in the z direction), and is magnetized in the +z direction. When a heat flow Q (∝-∇T) in the +x direction flows through the thermoelectric conversion element 1, a temperature difference occurs in the +x direction. As a result, an electromotive force V (∝M×(-∇T)) is generated in the thermoelectric conversion element 1 in the cross product direction (y direction) perpendicular to both the direction of the heat flow Q (+x direction) and the direction of magnetization M (+z direction) due to the anomalous Nernst effect.

[0022] Figure 3 shows the Nernst coefficient (S yx ) with the Nernst coefficient of thermoelectric conversion elements made of other metal materials, and shows the temperature dependence of the Nernst coefficient. Figure 4 is a graph in which the Nernst coefficients of the metal materials shown in Figure 3 are normalized by the Nernst coefficient at T = 300K.

[0023] In Figures 3 and 4, Fe3Al♯1 represents the observation results for thermoelectric conversion element 1 made of a single crystal with an off-stoichiometric composition (Fe-rich, Al-poor) in which the composition ratio of Fe and Al deviates from 3:1, and Fe3Al♯2 represents the observation results for thermoelectric conversion element 1 made of a single crystal with a stoichiometric composition of Fe and Al in which the composition ratio is 3:1.

[0024] In addition, in Figs. 3 and 4, L10 type MnGa, D0 22 The data for Mn2Ga, Co / Ni, FePd, and FePt are based on data disclosed in Appl. Phys. Lett. 106, 252405 (2015), and the data for Fe3O4 are based on data disclosed in Physical Review B 90, 054422 (2014). The data for Co2MnGa are based on research by the present inventors (Nature Physics 14, 1119-1124 (2018); International Publication No. 2019 / 009308).

[0025] From Figs. 3 and 4, the absolute values ​​of the Nernst coefficients |S yx | is larger than that of other metallic materials except Co2MnGa. In particular, |S of stoichiometric Fe3Al♯2 yx | is larger than that of off-stoichiometric Fe3Al♯1, and is about 4μV / K at room temperature (T = about 300K). yx It can be seen that the value is close to |≒6μV / K.

[0026] In addition, from Figs. 3 and 4, the S of off-stoichiometric Fe3Al♯1 yx is hardly affected by temperature changes in the temperature range of 200K to 400K, which includes room temperature, and is an almost constant value.

[0027] Figure 5 shows the Nernst coefficients (S yx) is shown in Fig. 5, and the temperature dependence of the Nernst coefficient in single crystals (S1, S2) and polycrystalline (P2) of Fe-Al alloy, and in polycrystalline Fe-Al-V alloy (P1) when a magnetic field B = 2 T is applied is shown in Fig. 6.

[0028] 5 and 6, S1 and S2 correspond to Fe3Al♯2 and Fe3Al♯1 shown in FIG. 3, respectively. S1 represents the observation results when a magnetic field B parallel to

[0001] is applied to the thermoelectric conversion element 1 and a heat flow Q parallel to

[0010] is caused to flow, and S2 represents the observation results when a magnetic field B parallel to

[0001] is applied to the thermoelectric conversion element 1 and a heat flow Q parallel to

[0210] is caused to flow.

[0029] In addition, in Figs. 5 and 6, P1 is Fe3Al in which part of the Fe site is replaced with vanadium (V). 2.8 V 0.15 P1 shows the observation results for thermoelectric conversion element 1 made of Al polycrystal, and P2 shows the observation results for thermoelectric conversion element 1 made of polycrystal with a stoichiometric composition of Fe and Al in a ratio of 3:1.

[0030] From Figures 5 and 6, the single crystals (S1, S2) have a larger absolute value of the Nernst coefficient |S yx In addition, when comparing the polycrystals P1 and P2, the stoichiometric binary polycrystal P2 has a larger |S than the ternary polycrystal P1. yx It can be seen that | is large.

[0031] In addition, the Nernst coefficient of the polycrystalline bodies (P1, P2) is almost unaffected by temperature changes in the temperature range of 200K to 400K, including room temperature (T = around 300K), and is almost constant, compared to the single crystalline bodies (S1, S2). yx Although | is small, |S yx is approximately 1.5 to 2.0 μV / K, it can be said that it has reached a practical level for use as a heat flow sensor, etc. Also, as described above, the polycrystals (P1, P2) are easier to fabricate than the single crystals (S1, S2).

[0032] Next, a thermoelectric conversion device in which the thermoelectric conversion element of this embodiment is modularized will be described. EXAMPLES

[0033] 7 shows the external configuration of a thermoelectric converter 20 according to Example 1 of this embodiment. The thermoelectric converter 20 includes a substrate 22 and a power generator 23 placed on the substrate 22. In the thermoelectric converter 20, when a heat flow Q flows from the substrate 22 side toward the power generator 23, a temperature difference occurs in the power generator 23 in the heat flow direction, and a voltage V is generated in the power generator 23 due to the anomalous Nernst effect.

[0034] The substrate 22 has a first surface 22a on which the power generating body 23 is placed, and a second surface 22b opposite to the first surface 22a. Heat from a heat source (not shown) is applied to the second surface 22b.

[0035] The power generating body 23 has a plurality of thermoelectric conversion elements 24 and a plurality of thermoelectric conversion elements 25, each of which has an L-shaped three-dimensional shape and is made of the same material as the thermoelectric conversion element 1 shown in Fig. 2. As shown in Fig. 7, the plurality of thermoelectric conversion elements 24 and the plurality of thermoelectric conversion elements 25 are alternately arranged in parallel on the substrate 22 in a direction (y direction) perpendicular to each longitudinal direction (x direction). The number of thermoelectric conversion elements 24 and thermoelectric conversion elements 25 constituting the power generating body 23 is not limited.

[0036] Moreover, the multiple thermoelectric conversion elements 24 and the multiple thermoelectric conversion elements 25 are arranged such that the direction of magnetization M1 of the thermoelectric conversion element 24 is opposite to the direction of magnetization M2 of the thermoelectric conversion element 25. Moreover, the multiple thermoelectric conversion elements 24 and the multiple thermoelectric conversion elements 25 have Nernst coefficients of the same sign.

[0037] The thermoelectric conversion element 24 has a first end surface 24a and a second end surface 24b parallel to the longitudinal direction (x direction). The thermoelectric conversion element 25 has a first end surface 25a and a second end surface 25b parallel to the longitudinal direction (x direction). The first end surface 25a of the thermoelectric conversion element 25 is connected to the second end surface 24b of the adjacent thermoelectric conversion element 24, and the second end surface 25b of the thermoelectric conversion element 25 is connected to the first end surface 24a of the adjacent thermoelectric conversion element 24 on the opposite side. This electrically connects the multiple thermoelectric conversion elements 24 and the multiple thermoelectric conversion elements 25 in series. That is, the power generation body 23 is provided in a serpentine shape on the first surface 22a of the substrate 22.

[0038] When heat is applied from a heat source to second surface 22b of substrate 22, heat flow Q flows in the +z direction toward power generator 23. When a temperature difference occurs due to heat flow Q, an electromotive force E1 is generated in thermoelectric conversion element 24 in a direction (-x direction) perpendicular to both the direction of magnetization M1 (-y direction) and the direction of heat flow Q (+z direction) due to the anomalous Nernst effect. In thermoelectric conversion element 25, an electromotive force E2 is generated in a direction (+x direction) perpendicular to both the direction of magnetization M2 (+y direction) and the direction of heat flow Q (+z direction) due to the anomalous Nernst effect.

[0039] As described above, the thermoelectric conversion elements 24 and 25 arranged in parallel are electrically connected in series, so that the electromotive force E1 generated in one thermoelectric conversion element 24 can be applied to the adjacent thermoelectric conversion element 25. In addition, since the electromotive force E1 generated in one thermoelectric conversion element 24 and the electromotive force E2 generated in the adjacent thermoelectric conversion element 25 are in the opposite directions, the electromotive forces of the adjacent thermoelectric conversion elements 24 and 25 are added together, and the output voltage V can be increased.

[0040] As a modified example of the thermoelectric conversion device 20 in FIG. 7, a configuration may be adopted in which adjacent thermoelectric conversion elements 24 and thermoelectric conversion elements 25 have Nernst coefficients of opposite signs, and the magnetization directions of the multiple thermoelectric conversion elements 24 and the multiple thermoelectric conversion elements 25 are the same (i.e., the direction of magnetization M1 and the direction of magnetization M2 are the same). EXAMPLES

[0041] FIG. 8 shows a plan view of a thermoelectric converter 20A according to Example 2 of this embodiment. The thermoelectric converter 20A has a plurality of rectangular parallelepiped thermoelectric converter elements 1A having the same sign as the Nernst coefficient as the power generator 23A. Each thermoelectric converter element 1A is made of the same material as the thermoelectric converter element 1 shown in FIG. 2. The plurality of thermoelectric converter elements 1A are arranged in parallel on the substrate 22A in a direction (y direction) perpendicular to the longitudinal direction (x direction) so that the directions of magnetization M are the same (y direction). The plurality of thermoelectric converter elements 1A are electrically connected in series by connecting adjacent thermoelectric converter elements 1A with copper wiring 26. The heat flow is caused to flow from the substrate 22A side toward the power generator 23A (z direction). The thermoelectric converter 20A has a configuration in which adjacent thermoelectric converter elements 1A are connected via copper wiring 26, and therefore can be manufactured more easily than the thermoelectric converter 20 of Example 1 shown in FIG. 7. EXAMPLES

[0042] In the thermoelectric mechanism based on the anomalous Nernst effect, the temperature gradient, the magnetization direction, and the voltage direction are perpendicular to each other, so it is possible to fabricate a thin sheet-like thermoelectric conversion element.

[0043] FIG. 9 shows the external configuration of a thermoelectric conversion device 30 according to a third embodiment, which includes a sheet-shaped thermoelectric conversion element 32. Specifically, the thermoelectric conversion device 30 includes a hollow member 31 and a long sheet-shaped (tape-shaped) thermoelectric conversion element 32 wound around the outer surface of the hollow member 31 so as to cover it. The thermoelectric conversion element 32 is made of the same material as the thermoelectric conversion element 1 shown in FIG. 2. The magnetization direction of the thermoelectric conversion element 32 is parallel to the longitudinal direction (x direction) of the hollow member 31. When a heat flow occurs in a direction perpendicular to the longitudinal direction (x direction) of the hollow member 31 and a temperature gradient occurs from the inside to the outside of the hollow member 31, a voltage V is generated along the longitudinal direction of the long thermoelectric conversion element 32 (a direction perpendicular to the magnetization direction and the heat flow direction) due to the anomalous Nernst effect.

[0044] In the thermoelectric converter 30 of FIG. 9, instead of the long sheet-like thermoelectric conversion element 32, a configuration in which a wire thermoelectric conversion element is wound around the hollow member 31 may be adopted.

[0045] 7 to 9, if the longitudinal length of the thermoelectric conversion element is L and the thickness (height) is H, the voltage generated by the anomalous Nernst effect is proportional to L / H. In other words, the longer and thinner the thermoelectric conversion element is, the greater the generated voltage will be. Therefore, by employing a power generator in which multiple thermoelectric conversion elements are electrically connected in series, or a wire or long sheet-like thermoelectric conversion element, it is expected that the anomalous Nernst effect will be improved.

[0046] The thermoelectric conversion devices shown in Examples 1 to 3 can be used in various applications. In particular, applications as a standalone power source or heat flow sensor for Internet of Things (IoT) sensors in the temperature range from room temperature to several hundreds of degrees Celsius are being considered.

[0047] For example, by applying the thermoelectric conversion device of this embodiment to a heat flow sensor, it is possible to judge whether the thermal insulation performance of a building is good or bad. In addition, by providing a thermoelectric conversion device in an exhaust system of an automobile or the like, it is possible to generate power using the heat (waste heat) of the exhaust gas, and the thermoelectric conversion device can be effectively used as an auxiliary power source. In addition, by arranging a heat flow sensor in a mesh shape on the wall surface of a certain space, it is possible to perform spatial recognition of heat flow and heat source. This is expected to be applied, for example, to high-precision temperature management for high-density agricultural crop cultivation and livestock growth, and as a driver detection system for automatic driving. Furthermore, the heat flow sensor can be used in indoor air conditioning management and deep body temperature management in medicine. In addition, by making the thermoelectric conversion element of this embodiment into a powder or a paste, it is expected to be applied to a wide range of fields.

[0048] In this embodiment, attention is focused on the voltage generated by the anomalous Nernst effect, but it is possible to increase the output voltage by the synergistic effect of the voltage caused by the Seebeck effect generated by the temperature gradient, the Hall effect generated based on the voltage created by the Seebeck effect, and the voltage generated by the anomalous Nernst effect.

[0049] Thus, according to the thermoelectric conversion element of this embodiment, the anomalous Nernst effect can be expressed by an alloy of Fe and Al, which is a material with a large Clarke number, is inexpensive, and is non-toxic. In particular, by adjusting the composition ratio of Fe and Al to adopt an off-stoichiometric composition, or by adopting a polycrystalline material rather than a single crystal, a thermoelectric conversion element can be provided that is insensitive to temperature changes over a wide range of Nernst coefficients from 200 K to 400 K. This eliminates the need to provide a temperature calibration circuit or a thermometer, which was required in a heat flow sensor or the like that uses a material whose Nernst coefficient changes significantly with temperature near room temperature, and makes it possible to make the thermoelectric conversion device more inexpensive.

[0050] 3 to 6, the thermoelectric conversion element is made of an Fe-Al alloy or an Fe-Al-V alloy in which part of the Fe sites of the Fe-Al alloy is replaced with V, but a transition element or typical element other than Al, or a transition element other than V may be used. That is, a first substance having a stoichiometric composition represented by Fe3X (X is a typical element or transition element), a second substance having an off-stoichiometric composition in which the composition ratio of Fe to X deviates from 3:1, a third substance in which part of the Fe sites of the first substance or part of the Fe sites of the second substance is replaced with a typical metal element or transition element other than X, a third substance having a composition formula of Fe3M1 1-x M2 xIn the fourth substance represented by (0 < x < 1) where M1 and M2 are different typical elements from each other, or in the fifth substance in which a part of the Fe sites of the first substance is replaced with a transition element other than X and a part of the sites of X is replaced with a typical metal element other than X, the manifestation of the anomalous Nernst effect can also be expected. As candidates for X other than Al, Ga, Ge, Sn, Si, Ti, Zr, Hf, V, Nb, Ta, Cr, Mo, W, Sc, Ni, Mn, or Co can be mentioned. As combinations of M1 and M2 constituting the fourth substance, for example, Ga and Al, Si and Al, Ga and B, etc. can be mentioned.

[0051] For example, the anomalous Nernst effect also appears in Fe-Ge alloys, Fe-Ga alloys, and Fe-Ga-Al alloys. In Fig. 10, Fe3Pt, Fe3Ge, Fe3Al, Fe3Ga 0.5 Al 0.5 and the magnetic field dependence of the Nernst coefficient (S yx ) of Fe3Ga and Co2MnGa at T = 300 K are shown, and in Fig. 11, the temperature dependence of the Nernst coefficient of Fe3Pt, Fe3Ge, Fe3Al, Fe3Ga, and Co2MnGa is shown.

[0052] In Figs. 10 and 11, Fe3Ge represents the observation results when a magnetic field B parallel to the a-axis is applied to the thermoelectric conversion element 1 composed of a hexagonal Fe3Ge single crystal and a heat current Q parallel to the c-axis is passed; Fe3Al represents the observation results when a magnetic field B parallel to

[0001] is applied to the thermoelectric conversion element 1 composed of a cubic Fe3Al single crystal and a heat current Q parallel to

[0110] is passed; Fe3Ga represents the observation results when a magnetic field B parallel to

[0110] is applied to the thermoelectric conversion element 1 composed of a cubic Fe3Ga single crystal and a heat current Q parallel to [1-11] is passed. Also, in Fig. 10, Fe3Ga 0.5 Al 0.5 is Fe3Ga 0.5 Al 0.510 and 11 show the observation results when a magnetic field B parallel to

[0110] is applied to a thermoelectric conversion element 1 made of a single crystal, and a heat flow Q parallel to [1-11] is passed through it. Furthermore, in Fig. 10 and Fig. 11, Fe3Pt shows the observation results when a magnetic field B parallel to

[0110] is applied to a thermoelectric conversion element 1 made of a Fe3Pt single crystal, and a heat flow Q parallel to [1-10] is passed through it, and Co2MnGa shows the observation results when a magnetic field B parallel to

[0001] is applied to a thermoelectric conversion element 1 made of a Co2MnGa single crystal, and a heat flow Q parallel to

[0110] is passed through it.

[0053] 10 and 11, the Fe3Ge single crystal has a higher |S yx Although | is smaller, |S yx | is large, and |S at room temperature yx | exceeds 2.0μV / K, which is sufficient for practical use in heat flow sensors, etc. 0.5 Al 0.5 The single crystal has a higher |S yx | is large, and is approximately 5.0μV / K at room temperature. Furthermore, |S yx | is well above 5.0 μV / K, approaching the highest value (6 μV / K) achieved to date for a Co2MnGa single crystal. The anomalous Nernst effect has also been confirmed in a hexagonal Fe3Ga single crystal.

[0054] Also, from FIG. 11, it is seen that the Nernst coefficients of the single crystals of Fe3Ge, Fe3Al, and Fe3Ga change more gradually with temperature in the temperature range of 200K to 400K than those of the single crystals of Fe3Pt and Co2MnGa.

[0055] The anomalous Nernst effect also appears in Fe-Sn alloys. Figure 12A shows the temperature dependence of the Nernst coefficient of a hexagonal Fe3Sn polycrystalline body obtained by melting and synthesizing the polycrystalline body in an arc furnace, then annealing it at 805 degrees Celsius for one week. Figure 12B shows the temperature dependence of the transverse thermoelectric conductivity α [A / Km] of this Fe3Sn estimated by the anomalous Nernst effect in a magnetic field B = 2T. These experimental data show that the Fe3Sn polycrystalline body has a Nernst coefficient |S yx It can be seen that | has increased to more than 3 μV / K.

[0056] Next, we will explain the results of producing polycrystals of a mixed crystal system between the three polycrystals (Fe3Si, Fe3Al, and Fe3Ga) and measuring the anomalous Nernst effect.

[0057] FIG. 13A and FIG. 13B show the mixed crystal system Fe3Si 1-x Al x The Nernst coefficient at T = 300K for the polycrystalline Fe3Si (0≦x≦1) shown in FIG. 0.67 Al 0.4 is a soft magnetic material called Sendust. From Fig. 13A and Fig. 13B, Fe3Si 0.3 Al 0.7 Except for the above, the Nernst coefficient |S yx It can be seen that | increases.

[0058] FIG. 14A and FIG. 14B show a mixed crystal system of Fe3Al and Fe3Ga. 1-x Ga x The Nernst coefficients at T = 300 K for the polycrystalline body (0 ≦ x ≦ 1) are shown in Fig. 14B. 1-x Ga x The Nernst coefficient of the single crystal is also shown. From Fig. 14A and Fig. 14B, the Nernst coefficient |S yx | is large, and both single crystals and polycrystals have Fe3Al 0.25 Ga 0.75 Except for polycrystalline materials, the Nernst coefficient |Syx It can be seen that | increases.

[0059] Figure 15 shows the Nernst coefficients at T = 300 K for single crystals and polycrystals (0.6 < x ≤ 1) of Fe3Cu obtained by substituting part of the Ga sites in Fe3Ga with Cu. 1-x Ga x The Nernst coefficients at T = 300 K for single crystals and polycrystals (0.6 < x ≤ 1) of Fe3CuGa are shown. The single crystals of Fe3CuGa are also those prepared at a crystal growth rate of 20 mm / h. From Figure 15, it can be seen that for the single crystals of Fe3CuGa, as the Cu content increases and the Ga content decreases, the Nernst coefficient |S| decreases. However, even at x = 0.7, it exceeds 3 μV / K, indicating that it has reached the practical level. 1-x Ga x The single crystals are also those prepared at a crystal growth rate of 20 mm / h. From Figure 15, for the single crystals of Fe3CuGa, as the Cu content increases and the Ga content decreases, the Nernst coefficient |S| decreases. However, even at x = 0.7, it exceeds 3 μV / K, indicating that it has reached the practical level. 1-x Ga x For the single crystals of Fe3CuGa, it can be seen that as the Cu content increases and the Ga content decreases, the Nernst coefficient |S| decreases. However, even at x = 0.7, it exceeds 3 μV / K, indicating that it has reached the practical level. yx It can be seen that | decreases. However, even at x = 0.7, it exceeds 3 μV / K, indicating that it has reached the practical level.

[0060] Next, referring to FIGS. 16A to 20B, the experimental results of doping Fe3Ga with Nd, Ho, Y, and Tb will be described.

[0061] Figure 16A shows the X-ray diffraction patterns of NdFeGa and Fe3Ga, and Figure 16B shows the magnetic field dependence of the magnetization of NdFeGa at T = 300 K. From Figure 16A, it can be seen that NdFeGa maintains almost the same crystal structure as Fe3Ga. Also, from Figure 16B, it can be seen that NdFeGa reaches a saturation magnetization of 4.08 μ / F.U. at 500 Oe, and the coercive force is hardly observable. 0.1 Fe 2.9 Figure 16A shows the X-ray diffraction patterns of NdFeGa and Fe3Ga, and Figure 16B shows the magnetic field dependence of the magnetization of NdFeGa at T = 300 K. From Figure 16A, it can be seen that NdFeGa maintains almost the same crystal structure as Fe3Ga. Also, from Figure 16B, it can be seen that NdFeGa reaches a saturation magnetization of 4.08 μ / F.U. at 500 Oe, and the coercive force is hardly observable. 0.1 Fe 2.9 Figure 16A shows the X-ray diffraction patterns of NdFeGa and Fe3Ga, and Figure 16B shows the magnetic field dependence of the magnetization of NdFeGa at T = 300 K. From Figure 16A, it can be seen that NdFeGa maintains almost the same crystal structure as Fe3Ga. Also, from Figure 16B, it can be seen that NdFeGa reaches a saturation magnetization of 4.08 μ / F.U. at 500 Oe, and the coercive force is hardly observable. 0.1 Fe 2.9 Figure 16A shows the X-ray diffraction patterns of NdFeGa and Fe3Ga, and Figure 16B shows the magnetic field dependence of the magnetization of NdFeGa at T = 300 K. From Figure 16A, it can be seen that NdFeGa maintains almost the same crystal structure as Fe3Ga. Also, from Figure 16B, it can be seen that NdFeGa reaches a saturation magnetization of 4.08 μ / F.U. at 500 Oe, and the coercive force is hardly observable. 0.1 Fe 2.9 Figure 16A shows the X-ray diffraction patterns of NdFeGa and Fe3Ga, and Figure 16B shows the magnetic field dependence of the magnetization of NdFeGa at T = 300 K. From Figure 16A, it can be seen that NdFeGa maintains almost the same crystal structure as Fe3Ga. Also, from Figure 16B, it can be seen that NdFeGa reaches a saturation magnetization of 4.08 μ / F.U. at 500 Oe, and the coercive force is hardly observable. B / F.U. and the coercive force is hardly observable.

[0062] Figure 16C shows the magnetic field dependence of the Nernst coefficients at T = 300 K for single crystals of Fe3Ga and polycrystals of NdFeGa. Also, Figure 16D shows the magnetic field dependence of the Hall resistivity at T = 300 K for single crystals of Fe3Ga and polycrystals of NdFeGa. From Figure 16C, for NdFeGa 0.1 Fe 2.9 Figure 16C shows the magnetic field dependence of the Nernst coefficients at T = 300 K for single crystals of Fe3Ga and polycrystals of NdFeGa. Also, Figure 16D shows the magnetic field dependence of the Hall resistivity at T = 300 K for single crystals of Fe3Ga and polycrystals of NdFeGa. From Figure 16C, for NdFeGa 0.1 Fe 2.9 Figure 16C shows the magnetic field dependence of the Nernst coefficients at T = 300 K for single crystals of Fe3Ga and polycrystals of NdFeGa. Also, Figure 16D shows the magnetic field dependence of the Hall resistivity at T = 300 K for single crystals of Fe3Ga and polycrystals of NdFeGa. From Figure 16C, for NdFeGa0.1 Fe 2.9 Nernst coefficient |S of Ga polycrystal yx Although the value of | is smaller than that of the Fe3Ga single crystal, it exceeds 3 μV / K, which is a practical level. 0.1 Fe 2.9 Hall resistivity ρ of Ga polycrystal yx It can be seen that the magnetic field dependence of shows almost the same behavior.

[0063] FIG. 17A shows the 0.05 Fe 2.95 The X-ray diffraction patterns of Ga and Fe3Ga are shown in FIG. 17B. 0.05 Fe 2.95 The magnetic field dependence of the magnetization of Ga at T = 300 K is shown in Fig. 17A. 0.05 Fe 2.95 It can be seen that Ga maintains almost the same crystal structure as Fe3Ga. Also, from FIG. 17B, Ho 0.05 Fe 2.95 Ga has a saturation magnetization of 4.8μ at 2T. B / FU, and almost no coercive force is observed.

[0064] In FIG. 18A, 0.05 Fe 2.95 The X-ray diffraction patterns of Ga and Fe3Ga are shown in FIG. 0.05 Fe 2.95 The magnetic field dependence of the magnetization of Ga is shown in Fig. 18A. 0.05 Fe 2.95 It can be seen that Ga maintains almost the same crystal structure as Fe3Ga. Also, from FIG. 18B, Y 0.05 Fe 2.95 Ga has a saturation magnetization of 3.28μ at 500 Oe. B / FU and has a coercive force of about 20 Oe.

[0065] Figure 19A shows the Tb 0.05 Fe 2.95 FIG. 19B shows an enlarged graph of the low magnetic field region of FIG. 19A.0.05 Fe 2.95 Ga has a saturation magnetization of 7.5μ at 2T. B / FU and has a coercive force of about 40 Oe.

[0066] Figure 20A shows the Tb 0.03 Fe 2.97 The X-ray diffraction patterns of Ga and Fe3Ga are shown in Fig. 20B. 0.03 Fe 2.97 The magnetic field dependence of the magnetization of Ga at T = 300K is shown in Fig. 20A. 0.03 Fe 2.97 It can be seen that Ga maintains almost the same crystal structure as Fe3Ga. Also, from FIG. 20B, Tb 0.03 Fe 2.97 Ga has a saturation magnetization of 6.9μ at 2T. B / FU, and almost no coercive force is observed.

[0067] 16A to 17B and 19A to 20B show the anomalous Nernst effect of a substance (third substance) obtained by substituting a part of the Fe site of Fe3Ga with Nd, Ho, or Tb, but a similar anomalous Nernst effect can be expected in other third substances obtained by substituting other lanthanides (e.g., Gd, etc.).

[0068] Next, with reference to Figs. 21A to 23B, experimental results in which Fe3Ga was doped with B, Mn, or Pt will be described.

[0069] First, the experimental results of doping Fe3Ga with B will be described. 0.8 B 0.2 , Fe3Ga 0.9 B 0.1 The X-ray diffraction patterns of Fe3Ga and Fe3Ga are shown. 0.8 B 0.2 and Fe3Ga 0.9 B 0.1It can be seen that the Fe3Ga alloy maintains a crystal structure almost identical to that of Fe3Ga. In addition, energy dispersive X-ray analysis (EDX) (not shown) shows that the additive B appears near the boundary of Fe3Ga.

[0070] In FIG. 21B, Fe3Ga 0.8 B 0.2 The magnetic field dependence of magnetization at T=300K is shown for a needle-shaped sample and a plate-shaped sample consisting of the above. The needle-shaped sample (cylindrical sample) shown in Fig. 21B is magnetized in the longitudinal direction, and a magnetic field is applied parallel to the longitudinal direction. The plate-shaped sample shown in Fig. 21B is magnetized in the in-plane direction, and a magnetic field is applied perpendicular to the plane. As described above, since the magnetization direction and the magnetic field direction are perpendicular to each other in the plate-shaped sample, a strong magnetic field must be applied to make the magnetization rise to the magnetic field direction. From Fig. 21B, it can be seen that when a strong magnetic field is applied, the magnetization of the plate-shaped sample changes only slightly linearly, but the magnetization of the needle-shaped sample, which has magnetization parallel to the magnetic field, changes significantly, and clear hysteresis appears.

[0071] Fig. 21C shows the needle-shaped sample of Fe3Ga 0.8 B 0.2 The magnetic field dependence of the magnetization of Fe3Ga and Fe3Ga is shown in Fig. 21C. The needle-shaped sample shown in Fig. 21C is magnetized in the longitudinal direction, and a magnetic field is applied parallel to the magnetization direction. Since the magnetization direction and the magnetic field direction are parallel, both Fe3Ga and Fe3Ga 0.8 B 0.2 The magnetization is saturated at relatively weak magnetic fields (approximately 400 Oe and 800 Oe, respectively). 0.8 B 0.2 The hysteresis is more obvious in Fe3Ga than in Fe3Ga, and the coercive force is about 10 Oe in Fe3Ga. 0.8 B 0.2 The coercive force is approximately 35 Oe.

[0072] Figure 21D shows the Fe3Ga plate specimen. 0.8 B 0.2The magnetic field dependence of magnetization of Fe3Ga and Fe3Ga is shown in Fig. 21D. The plate-shaped sample shown in Fig. 21D is magnetized in the in-plane direction, and a magnetic field is applied in the perpendicular direction. Since the magnetization direction and the magnetic field direction are perpendicular to each other, a strong magnetic field is required to raise the magnetization in the perpendicular direction, and both Fe3Ga and Fe3Ga 0.8 B 0.2 Even for Fe3Ga, the magnetization increases linearly up to a magnetic field exceeding 3K e, and the hysteresis is weak. In fact, almost no coercivity is observed for Fe3Ga. On the other hand, the enlarged view near the low magnetic field (insert in Fig. 21D) shows that Fe3Ga 0.8 B 0.2 has a coercivity of about 35 Oe.

[0073] Figure 21E shows the Fe3Ga plate sample. 0.8 B 0.2 The magnetic field dependence of the Nernst coefficient of the plate-shaped sample Fe3Ga is shown in Fig. 21F. 0.8 B 0.2 21E and 21F show the magnetic field dependence of Hall resistivity of Fe3Ga 303N ... 0.8 B 0.2 Nernst coefficient of |S yx reaches 4 μV / K, which is approximately 80% of the Nernst coefficient of the Fe3Ga polycrystalline body (4.9 μV / K: see FIGS. 14A and 14B).

[0074] In this way, by substituting a part of the Ga site of Fe3Ga with B, the coercive force is increased and the Nernst coefficient can be secured to a value of about 80% of that of Fe3Ga. 0.8 B 0.2 This can be said to be advantageous for fabricating a thermopile that can be realized in a zero magnetic field.

[0075] Next, the experimental results of doping Fe3Ga with Mn will be described. 2.9 Mn0.1 Ga, Fe 2.5 Mn 0.5 The X-ray diffraction patterns of Ga, Fe2MnGa, and Fe3Ga are shown in Fig. 22B. 2.9 Mn 0.1 Ga and Fe 2.5 Mn 0.5 22B shows the magnetic field dependence of magnetization of Ga at T=300 K. The needle-shaped sample shown in FIG. 22B is magnetized in the longitudinal direction, and a magnetic field is applied parallel to the magnetization direction.

[0076] From Fig. 22A, Fe 2.9 Mn 0.1 Ga and Fe 2.5 Mn 0.5 It can be seen that Ga maintains almost the same crystal structure as Fe3Ga. Also, from FIG. 22B, Fe 2.9 Mn 0.1 Although Ga shows a large magnetization, it shows almost no hysteresis. 2.5 Mn 0.5 Ga shows small hysteresis and has a coercivity of about 10 Oe (same as the coercivity of the needle-shaped Fe3Ga sample shown in Figure 21C), while Fe 2.9 Mn 0.1 It can be seen that the magnetization is significantly suppressed compared to Ga.

[0077] Next, the experimental results of doping Pt into Fe3Ga are explained. 2.9 Pt 0.1 Ga, Fe 2.9 Pt 0.1 Ga 0.9 Ge 0.1 FIG. 23B shows the X-ray diffraction patterns of the needle-shaped sample Fe 2.9 Pt 0.1 Ga and Fe 2.9 Pt 0.1 Ga 0.9 Ge 0.1 23B shows the magnetic field dependence of magnetization at T=300 K. The needle-shaped sample shown in FIG. 23B is magnetized in the longitudinal direction, and a magnetic field is applied parallel to the magnetization direction.

[0078] From Fig. 23A, Fe2.9 Pt 0.1 Ga and Fe 2.9 Pt 0.1 Ga 0.9 Ge 0.1 It can be seen that Fe3Ga maintains a crystal structure almost similar to that of Fe3Ga. 2.9 Pt 0.1 Ga shows almost no hysteresis. 2.9 Pt 0.1 Ga 0.9 Ge 0.1 exhibits small hysteresis and has a coercivity of about 8 Oe. 2.9 Pt 0.1 It can be seen that the coercivity is increased by substituting some of the Ga sites with Ge.

[0079] As described above, an increase in coercivity was observed when some of the Ga sites of Fe3Ga were substituted with B (see FIGS. 21C and 21D), which suggests that the substitution of the Ga sites affects the magnetic properties.

[0080] 24 to 30, an embodiment in which the thermoelectric conversion element 1 is a thin film will be described. In the following embodiment, the thermoelectric conversion element 1 is made of a first substance having a composition formula Fe3X or a second substance having an off-stoichiometric composition in which the composition ratio of Fe and X is shifted from that of the first substance, but is not limited thereto.

[0081] First, a method for producing a thin film will be described. In the following, an example of producing a thin film by a sputtering method will be shown, but the method for producing a thin film is not limited thereto, and for example, a molecular beam epitaxy (MBE) method, a chemical vapor deposition (CVD), a pulsed laser deposition (PLD), a plating method, etc. may be adopted.

[0082] For example, in the preparation of a Fe3Ga thin film, a target with a composition of Fe and Ga in a ratio of 3:1 is discharged at room temperature in a DC magnetron sputtering device to produce a thin film of Fe doped with Ga on a magnesium oxide (MgO) substrate with a

[0001] orientation. The Fe3Ga sample formed at room temperature is a polycrystalline thin film, but by annealing the film at 500°C for 30 minutes without breaking the vacuum after deposition, an epitaxial thin film with a

[0001] orientation can be produced. In this way, the Fe3Ga thin film is an epitaxial thin film with a

[0001] orientation formed on a MgO substrate with a

[0001] orientation.

[0083] In addition, an oxidation prevention layer made of MgO is provided on the outermost surface of the thin film. Note that, as the oxidation prevention layer, a cap layer that prevents general oxidation, such as Al, Al2O3, SiO2, etc., other than MgO can also be used. Note that the buffer layer between the thin film and the substrate and the cap layer on the outermost surface of the thin film are not necessarily required.

[0084] In the above-mentioned manufacturing method, the T / S distance (distance between the target and the substrate) is preferably 15 cm to 20 cm in a sputtering apparatus, but even when other manufacturing methods are used, the T / S distance may be in the range of 5 cm to 40 cm.

[0085] It is difficult to obtain a voltage in a zero magnetic field with a ferromagnetic material because it is difficult to align the magnetization direction perpendicular to the temperature difference due to the effect of the demagnetizing field. However, when the film is thinned as described above, the contribution of the demagnetizing field in the direction perpendicular to the film's surface becomes large, while the effect of the demagnetizing field becomes almost zero in the in-plane direction. This stabilizes the magnetization in the in-plane direction.

[0086] Next, with reference to Fig. 24, a method for measuring the anomalous Nernst effect when a temperature gradient is applied in the in-plane direction to a thin film sample as a thermoelectric conversion element 1 will be described. In order to measure such an anomalous Nernst effect, for example, a rectangular parallelepiped structure as shown in Fig. 24 is used. In the structure of Fig. 24, a thin film (Fe3X thin film) sample having a thickness of 50 nm is laminated on a MgO substrate having a thickness of 500 µm, and a MgO cap layer having a thickness of 5 nm is laminated on the thin film sample. The length of this structure in the longitudinal direction is 9 mm, and the width is 2 mm. Thermocouples are provided in the thin film sample in the longitudinal direction, and the interval between the thermocouples is 6 mm.

[0087] As mentioned above, the thin film sample is magnetized in the in-plane direction. As shown in Figure 24, when a magnetic field is applied perpendicular to the surface of the thin film sample, the magnetization rises in the perpendicular direction. When a heat flow Q flows in the longitudinal direction of the thin film sample, a temperature difference ΔT (= T2 - T1) is generated. As a result, as in Figure 2, an electromotive force V is generated in a direction perpendicular to both the direction of the heat flow Q and the direction of magnetization (perpendicular to the surface) due to the anomalous Nernst effect. yx occurs.

[0088] The measurement results of the anomalous Nernst effect when the thin film sample shown in FIG. 24 is Fe3Ga (magnetic field dependence of the Nernst coefficient at T=300K) are shown in FIG. 25. From FIG. 25, the Nernst coefficient |S yx It can be seen that | indicates 4.0 μV / K.

[0089] 24 and 25 show examples in which a temperature gradient is applied in the in-plane direction to a thin-film sample, but the anomalous Nernst effect can also be obtained when a temperature gradient is applied in the perpendicular direction to the surface of a thin-film sample, as shown in Fig. 26. That is, the thin-film sample serving as a thermoelectric conversion element 1 is magnetized in the in-plane direction (x direction), and when a heat flow Q is applied to this thin-film sample in the perpendicular direction to the surface (z direction), an electromotive force V is generated in a direction (y direction) perpendicular to both the direction of the heat flow Q and the direction of magnetization M.

[0090] In order to measure such an anomalous Nernst effect, for example, a structure shown in FIG. 27A is used. In the structure of FIG. 27A, a 500 μm thick silicone pad is provided on a heat sink made of Cu, a 500 μm thick MgO substrate is laminated on the silicone pad, a 50 nm thick thin film sample is laminated on the MgO substrate, and a 5 nm thick MgO cap layer is laminated on the thin film sample. Voltage terminals are provided on both ends in the longitudinal direction of the MgO cap layer, and these voltage terminals are connected to both ends in the longitudinal direction of the thin film sample. As a result, as shown in FIG. 27B, an electromotive force V yx can be measured.

[0091] A 500 μm thick silicone pad is laminated on the MgO cap layer, a 1 mm thick copper plate is laminated on the silicone pad, and a resistive heater (ceramic heater) is provided on the copper plate. Thermocouples are provided at the top end of the MgO cap layer and the bottom end of the MgO substrate, and the temperature gradient ΔT in the perpendicular direction [00-1] generated by the heat flow from the ceramic heater from the top end of the MgO cap layer through the thin film sample to the bottom end of the MgO substrate is measured using the thermocouples. all In Figures 27A and 27B, the magnetization direction of the thin film sample and the direction of the applied magnetic field

[0110] are parallel.

[0092] Figure 28 shows the measurement results (magnetic field dependence of electromotive force) of the anomalous Nernst effect at 300K when the thin film sample shown in Figures 27A and 27B is Fe3Ga. Figure 28 shows that, unlike the bulk sample, the thin film sample Fe3Ga has an electromotive force of 19.8μV even in zero magnetic field, which is almost the same as the value in the saturated magnetic field. The coercive force of the thin film sample Fe3Ga is about 40 Oe.

[0093] In this measurement, the temperature difference measured by the thermocouple in the direction perpendicular to the surface is ΔT all = 1.5K. Here, the temperature gradient of the thin film itself, ∇T filmSince it is difficult to accurately estimate , the measurement results shown in Figure 25 are used, that is, the measurement results when a temperature gradient is applied in the in-plane direction of the thin film sample and a magnetic field is applied in the perpendicular direction. The Nernst coefficient (4.0 μV / K) of Fe3Ga shown in Figure 25 is ∇T film Assuming that this method is also applicable to the thin film sample Fe3Ga, the temperature gradient in the direction perpendicular to the surface of the 50 nm thin film sample Fe3Ga can be estimated to be 0.9 K / mm.

[0094] In the above-mentioned thin film preparation method, annealing is performed after film formation at room temperature. However, even in the case of polycrystalline or amorphous thin films obtained without annealing after film formation at room temperature, the same degree of anomalous Nernst effect can be obtained as in the case of annealed thin film samples, as will be shown below.

[0095] Thin film production is easier without annealing, and it can be used for flexible films. In addition, while the above-mentioned epitaxial film production requires a MgO substrate with a

[0001] orientation, the production of polycrystalline or amorphous films does not require any substrate.

[0096] The material of the substrate is not limited, and other than MgO, Si, Al2O3, PET, polyimide, etc. can be used.

[0097] Fig. 29A shows the measurement results (magnetic field dependence of electromotive force) of the anomalous Nernst effect at T = 300K when a temperature gradient is applied in the perpendicular direction as shown in Fig. 26 to a thin film sample Fe3Ga formed on an MgO substrate without annealing after film formation at room temperature, and Fig. 29B shows an enlarged graph of the low magnetic field area of ​​Fig. 29A. In Figs. 29A and 29B, as in Figs. 27A and 27B, the thickness of the thin film sample Fe3Ga is 50 nm, and the temperature gradient in the perpendicular direction measured by a thermocouple is ΔT all =1.5K(3K / mm).

[0098] 30 shows the measurement results of the anomalous Nernst effect (magnetic field dependence of electromotive force) when a temperature gradient is applied in the perpendicular direction for an epitaxial film of Fe3Ga formed at room temperature and then annealed, and an amorphous film of Fe3Ga formed at room temperature and then not annealed. The measurement results of the amorphous film in FIG. 30 correspond to the measurement results shown in FIG. 29A and FIG. 29B, and the measurement results of the epitaxial film in FIG. 30 correspond to the measurement results of Fe3Ga shown in FIG. 28.

[0099] 29A, 29B, and 30 show that the electromotive force of the amorphous film is almost equal to that of the epitaxial film under a high magnetic field, and is about half that of the epitaxial film under zero magnetic field. In other words, the Fe3Ga amorphous thin film sample fabricated by room temperature deposition has a Nernst coefficient of about 2μV / K under zero magnetic field.

[0100] In the above-described embodiment (FIGS. 24 to 30), the thin film has a thickness of 50 nm, but the thickness is not limited and may be 10 μm or less, and more preferably 1 μm or less.

[0101] As described above, according to the embodiment of Figs. 24 to 30, since a thin film made of Fe3Ga is used as the thermoelectric conversion element 1, the anomalous Nernst effect can be obtained with an inexpensive material. Furthermore, by making the thermoelectric conversion element 1 thin, it is possible to fabricate a thermoelectric conversion element 1 that exhibits a huge anomalous Nernst effect even in a zero magnetic field. Furthermore, the anomalous Nernst effect can be obtained whether the thermoelectric conversion element 1 is a single crystal, polycrystalline, or amorphous body. Furthermore, since the film can be formed at room temperature, a thin film can be formed on a flexible substrate that is vulnerable to heat. [Explanation of symbols]

[0102] 1, 1A, 24, 25, 32 Thermoelectric conversion element 20, 20A, 30 Thermoelectric conversion device 22, 22A board 23, 23A Power generating unit 31 Hollow members

Claims

1. A substrate, a thermoelectric conversion thin film formed above the substrate; an anti-oxidation layer formed above the thermoelectric conversion thin film; Equipped with The thermoelectric conversion thin film is A single-layer thin film made of Fe 3 Al with a thickness of less than 1 μm, The Fe 3 Al has at least a crystal structure of a DO 3 phase derived from a body-centered cubic lattice, outputting a voltage due to the anomalous Nernst effect based on a temperature difference in a direction parallel to a stacking direction of the thermoelectric conversion thin film relative to the substrate; Thermoelectric conversion device.

2. The thermoelectric conversion device according to claim 1, wherein the Nernst coefficient of the Fe 3 Al constituting the thermoelectric conversion thin film is constant in the temperature range of 200K to 400K.

3. The thermoelectric conversion device according to claim 1, wherein the Fe 3 Al crystals constituting the thermoelectric conversion thin film are single crystals.

4. The thermoelectric conversion device according to claim 1, wherein the Fe 3 Al crystals constituting the thermoelectric conversion thin film are polycrystalline.

5. A thermoelectric conversion device as described in claim 1, wherein a buffer layer is provided between the substrate and the thermoelectric conversion thin film.

6. The thermoelectric conversion device according to claim 1, wherein the anti-oxidation layer is composed of at least one of MgO, Al, Al 2 O 3 , and SiO 2 .

7. The thermoelectric conversion thin film is provided in plurality on the substrate, each of the plurality of thermoelectric conversion thin films has a shape extending in one direction; The thermoelectric conversion device according to claim 1 , wherein the plurality of thermoelectric conversion thin films are arranged in parallel in a direction perpendicular to the one direction and are electrically connected in series.

8. The thermoelectric conversion device according to claim 7 , wherein the plurality of thermoelectric conversion thin films are arranged in a serpentine pattern.

9. The substrate is cylindrical, The thermoelectric conversion device according to claim 1 , wherein the thermoelectric conversion thin film is provided on the outside of the cylindrical substrate so as to cover the surface.

Images