# Fileset

[Nature Communications--Decoupled charge and heat transport in Fe2VAl composite thermoelectrics with topological-insulating grain boundary networks.pdf](https://mdr.nims.go.jp/filesets/91b4399e-4042-4341-a0a2-0ee491d3a524/download)

## Creator

Fabian Garmroudi, [Illia Serhiienko](https://orcid.org/0000-0002-3072-9412), Michael Parzer, Sanyukta Ghosh, Pawel Ziolkowski, Gregor Oppitz, Hieu Duy Nguyen, [Cédric Bourgès](https://orcid.org/0000-0001-9056-0420), [Yuya Hattori](https://orcid.org/0000-0002-3805-4659), Alexander Riss, Sebastian Steyrer, Gerda Rogl, Peter Rogl, Erhard Schafler, [Naoyuki Kawamoto](https://orcid.org/0000-0002-2022-3987), Eckhard Müller, Ernst Bauer, Johannes de Boor, [Takao Mori](https://orcid.org/0000-0003-2682-1846)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Decoupled charge and heat transport in Fe2VAl composite thermoelectrics with topological-insulating grain boundary networks](https://mdr.nims.go.jp/datasets/14a4793a-86f7-4e81-8a27-d2a48eee7181)

## Fulltext

Decoupled charge and heat transport in Fe2VAl composite thermoelectrics with topological-insulating grain boundary networksArticle https://doi.org/10.1038/s41467-025-57250-6Decoupled charge and heat transport inFe2VAl composite thermoelectrics withtopological-insulating grain boundarynetworksFabian Garmroudi 1,2 , Illia Serhiienko 2, Michael Parzer 1,Sanyukta Ghosh3, Pawel Ziolkowski 3, Gregor Oppitz3, Hieu Duy Nguyen4,Cédric Bourgès 2,5, Yuya Hattori 2, Alexander Riss 1, Sebastian Steyrer1,Gerda Rogl 6, Peter Rogl6, Erhard Schafler7, Naoyuki Kawamoto4,EckhardMüller 3,8, Ernst Bauer1, Johannes de Boor 3,9 & TakaoMori 2,10Decoupling charge and heat transport is essential for optimizing thermo-electric materials. Strategies to inhibit lattice-driven heat transport, however,also compromise carrier mobility, limiting the performance of most thermo-electrics, including Fe2VAl Heusler compounds. Here, we demonstrate aninnovative approach, which bypasses this tradeoff: via liquid-phase sintering,we incorporate the archetypal topological insulator Bi1−xSbx betweenFe2V0.95Ta0.1Al0.95 grains. Structural investigations alongside extensive ther-moelectric and magneto-transport measurements reveal distinct modifica-tions in the microstructure, a reduced lattice thermal conductivity and asimultaneously enhanced carriermobility arising from topologically protectedcharge transport along the grain boundaries. This yields a huge performanceboost, resulting in oneof the highestfigure ofmerits amongboth half- and full-Heusler compounds, z ≈ 1.6 × 10−3 K−1 (zT ≈ 0.5) at 295 K. Our findings highlightthe potential of topological-insulating secondary phases to decouple chargeand heat transport and call for more advanced theoretical studies of multi-phase composites.Given the increasing global demand for efficient energy utilization,thermoelectrics (TEs) present a promising solution as they can harvestdecentralizedwaste heat sources or function as Peltier coolers, e.g., forthermal management applications1,2. The conversion efficiency of TEdevices depends on the hot- and cold-side temperatures and amaterial-dependent figure ofmerit, z∝ μW/κL. The highest achievable zin a semiconductor with optimized carrier concentration is deter-mined by the weighted carriermobility μWof electrons or holes, whichshould bemaximized, and by the lattice thermal conductivity κL, whichshould be minimized3,4. The inherent tradeoff between μW and κLReceived: 8 November 2024Accepted: 17 February 2025Check for updates1Institute of Solid State Physics, TUWien, Vienna, Austria. 2International Center for Materials Nanoarchitectonics (WPI-MANA), National Institute for MaterialsScience (NIMS), Tsukuba, Japan. 3Institute of Materials Research, German Aeropspace Center (DLR), Cologne, Germany. 4Center for Basic Research onMaterials (CBRM), National Institute for Materials Science (NIMS), Tsukuba, Japan. 5International Center for Young Scientists, National Institute for MaterialsScience (NIMS), Tsukuba, Japan. 6Institute of Materials Chemistry, University of Vienna, Vienna, Austria. 7Faculty of Physics, University of Vienna,Vienna, Austria. 8Institute of Inorganic andAnalytical Chemistry, Justus LiebigUniversityGiessen,Giessen,Germany. 9University of Duisburg-Essen, Faculty ofEngineering, Instituteof Technology forNanostructures (NST) andCENIDE,Duisburg,Germany. 10GraduateSchool of Pure andAppliedSciences, University ofTsukuba, Tsukuba, Japan. e-mail: f.garmroudi@gmx.at; Johannes.deBoor@dlr.de; mori.takao@nims.go.jpNature Communications |         (2025) 16:2976 11234567890():,;1234567890():,;http://orcid.org/0000-0002-0088-1755http://orcid.org/0000-0002-0088-1755http://orcid.org/0000-0002-0088-1755http://orcid.org/0000-0002-0088-1755http://orcid.org/0000-0002-0088-1755http://orcid.org/0000-0002-3072-9412http://orcid.org/0000-0002-3072-9412http://orcid.org/0000-0002-3072-9412http://orcid.org/0000-0002-3072-9412http://orcid.org/0000-0002-3072-9412http://orcid.org/0000-0003-3509-7474http://orcid.org/0000-0003-3509-7474http://orcid.org/0000-0003-3509-7474http://orcid.org/0000-0003-3509-7474http://orcid.org/0000-0003-3509-7474http://orcid.org/0000-0003-1519-6803http://orcid.org/0000-0003-1519-6803http://orcid.org/0000-0003-1519-6803http://orcid.org/0000-0003-1519-6803http://orcid.org/0000-0003-1519-6803http://orcid.org/0000-0001-9056-0420http://orcid.org/0000-0001-9056-0420http://orcid.org/0000-0001-9056-0420http://orcid.org/0000-0001-9056-0420http://orcid.org/0000-0001-9056-0420http://orcid.org/0000-0002-3805-4659http://orcid.org/0000-0002-3805-4659http://orcid.org/0000-0002-3805-4659http://orcid.org/0000-0002-3805-4659http://orcid.org/0000-0002-3805-4659http://orcid.org/0000-0002-9707-8394http://orcid.org/0000-0002-9707-8394http://orcid.org/0000-0002-9707-8394http://orcid.org/0000-0002-9707-8394http://orcid.org/0000-0002-9707-8394http://orcid.org/0000-0002-8056-5006http://orcid.org/0000-0002-8056-5006http://orcid.org/0000-0002-8056-5006http://orcid.org/0000-0002-8056-5006http://orcid.org/0000-0002-8056-5006http://orcid.org/0000-0001-9499-9473http://orcid.org/0000-0001-9499-9473http://orcid.org/0000-0001-9499-9473http://orcid.org/0000-0001-9499-9473http://orcid.org/0000-0001-9499-9473http://orcid.org/0000-0002-1868-3167http://orcid.org/0000-0002-1868-3167http://orcid.org/0000-0002-1868-3167http://orcid.org/0000-0002-1868-3167http://orcid.org/0000-0002-1868-3167http://orcid.org/0000-0003-2682-1846http://orcid.org/0000-0003-2682-1846http://orcid.org/0000-0003-2682-1846http://orcid.org/0000-0003-2682-1846http://orcid.org/0000-0003-2682-1846http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-57250-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-57250-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-57250-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-57250-6&domain=pdfmailto:f.garmroudi@gmx.atmailto:Johannes.deBoor@dlr.demailto:mori.takao@nims.go.jpwww.nature.com/naturecommunicationspresents one of the most formidable challenges in the design andoptimization of TE materials, requiring the decoupling of charge andheat transport, that is, the realization of a phonon-glass electron-crystal concept.Since the mid-20th century, Bi2Te3-based systems have been thegold standard for TEs operating near room temperature, and cur-rently, they remain the only commercially available option5,6. However,the scarcity of tellurium, along with the brittle nature and poormechanical properties of thesematerials limits their widespread use ineveryday life and industrial applications. Therefore, it is crucial toexplore alternatives thatoffer competitive performanceandovercomethe challenges related to Bi2Te3.For n-type materials, cost-effective Mg3(Bi,Sb)2 Zintl compoundshave been considered the hottest candidates as they exhibit very highz7–9. However, these materials, especially the Bi-rich alloys withattractive near-room temperature properties, suffer from poor che-mical stability and degrade rapidly when exposed to air, presenting anongoing challenge for practical applications.On the other hand, Heusler compounds based on Fe2VAl, thefocus of this study, exhibit excellent chemical andmechanical stability.These materials are also composed of earth-abundant, inexpensiveelements with great recyclability10 – sustainability aspects that arebecoming increasingly important worldwide, and particularly withinthe EU. Moreover, they display outstanding electronic transportproperties, with weighted mobilities that are comparable to or evengreater than those of other state-of-the-art TEs11. Yet, their intrinsicallylarge κL limits their potential as TE materials12. Consequently, previousstudies have primarily focused on reducing κL by substituting heavyelements13–15, lowering the dimensionality through thin-filmdeposition16–18, or reducing the grain size12,19,20. Although these strate-gies have resulted in enhancements of z, the overall performanceremains a significant bottleneck and is too low for most practicalapplications.In this study, we demonstrate that by incorporating chemicallyand structurally distinct Bi1−xSbx at the grain boundaries, charge,and heat transport can be decoupled, resulting in a reduction ofκL, and simultaneously, in an unexpected increase of μW (Fig. 1a).Consequently, the figure of merit is boosted by more than a factor oftwo, up to zmax � 1:7 × 10�3 K−1 at 240–250K (z ≈ 1.6 × 10−3 K−1 at roomtemperature), representing one of the largest values hitherto reportedamong n-type half- and full-Heusler compounds (Fig. 1b).ResultsDecoupling charge and heat transport in Fe2VAlThe lattice thermal conductivity of Fe2VAl Heusler compounds isintrinsically large, κL ≈ 27Wm−1 K−1 at 300K21, which can be mainlyattributed to a lack of structural and chemical bonding complexity aswell as the absence of heavy elements, leading to high sound velocities.Upon alloying, κL can be drastically reduced down to 10Wm−1 K−1 inFe2VAl1−xSix21, 7Wm−1 K−1 in Fe2VAl1−xGex21, and by substituting heavy5d elements, further down to 5Wm−1 K−1 in Fe2VTaxAl1−x14 and4Wm−1 K−1 in Fe2V1−xWxAl15. As a downside, the very same pointdefects, which effectively inhibit heat transport by high-frequencyphonons, also strongly scatter charge carriers. This is particularly truefor the 5d elements like Ta and W, which are substituted for V atoms.SinceV-d states dominate the electronic states of the conduction band,introducing substitutional disorder at the V site results in intenseelectronic scattering. We gathered TE property data from varioussubstitution studies13,14,21–23 and calculated μW4. A strong tradeoff rela-tionship between μW and κL is obvious from Fig. 1a (black symbols).Aside from introducing point defects, κL can be suppressed byreducing the grain size d and several studies attempted to enhance z bygrain size reduction, e.g., via ball milling13,19 or high-pressure torsion(HPT)20,24, yielding d ≈ 100nm. Employing HPT, Fukuta et al. recentlyreported very low values of κL down to 1.3Wm−1 K−1 in Fe2V0.98Ta0.1Al0.92at 350K and zT up to 0.37 at 400 K (z ≈ 0.9 × 10−3 K−1)24. These remark-able findings motivated us to (i) reproduce them and (ii) apply HPT to avariety of different samples with optimized compositions. The results ofthese endeavors are summarized in the Supplementary Information (SI).While κL could indeed be dramatically reduced down to <2Wm−1 K−1, weconcomitantly observed a huge deterioration of electronic transport inall cases (see Figs. S1 and S2 and blue symbols in Fig. 1a), resulting in noenhancement of zT (Fig. S3). Similar observations have been made, e.g.,for Mg3(Bi,Sb)225,26 and various half-Heuslers, where reducing grain sizecomes at a cost of reducing μW27,28. The discrepancy between our resultsand previous ones from ref. 29 suggests that setup-specific conditionsduring HPT are generally very important, complicating reproducibilityand upscale production.Fig. 1 | Boosting thermoelectric performance in Heusler compounds bydecoupled charge andheat transport. aTradeoff betweenweightedmobility andlattice thermal conductivity (plus bipolar term κB) in Fe2VAl-based Heusler com-pounds at room temperature13,14,21–23. Data for state-of-the-art n-type Bi2Te3- andMg3Bi2-based systems6,7 at 300K are shown for comparison. Composites in thiswork are found to bypass this tradeoff.bTemperature-dependent figureofmerit ofthe best composite from this work (FVAB50), compared to other optimally doped,high-performance n-type thermoelectrics50–55, reaching one of the highest z amongboth half-Heusler (hH) and full-Heusler (fH) compounds. c Schematic synthesis ofcomposites via liquid-phase sintering.Article https://doi.org/10.1038/s41467-025-57250-6Nature Communications |         (2025) 16:2976 2www.nature.com/naturecommunicationsInstead, we have devised a different approach wherein chemi-cally and structurally distinct Bi1−xSbx is incorporated as a secondaryphase between the Heusler grains. Figure 1c outlines the synthesisprocedure. The starting materials were first synthesized using aninduction melting furnace and then hand-ground into a fine powder.The powders were mixed in various ratios (5–50 vol.% Bi0.9Sb0.1), andsintered at 1373 K. The much lower melting point of Bi0.9Sb0.1 causesexcess liquid to be expelled during sintering. The retention ofBi1−xSbx in the composite depends on the particle size of the Heuslerphase and the amount of Bi0.9Sb0.1 used. Backscattered scanningelectron microscopy (BS-SEM) shows that up to ≈30 vol.%, Bi0.9Sb0.1fills only the triple junctions of Heusler grains, while 50 vol.%Bi0.9Sb0.1 allows the liquid phase to wet and coat all grains as a grainboundary (GB) phase (Fig. S10). This produces highly dense(Fe2V0.95Ta0.1Al0.95 + Bi0.9Sb0.1) composites, referred to as FVABX,with X indicating the Bi0.9Sb0.1 volume percentage added beforesintering. The reference sample, without any Bi1−xSbx, achieved adensity of approximately 95% of its theoretical density.When about 10 vol.% Bi0.9Sb0.1 are added before sintering, thecomposite material shows minimal porosity. SEM micrographs con-firm that any pores present in the initial sample are filled by thesecondary Bi1−xSbx phase, resulting in a density close to 100% for allcomposite samples.The μW versus κL trend (see Fig. 1a) for the composite samples isunusual and cardinally different from other approaches, like alloyingor grain size reduction via HPT. Moreover, the exceptionally high μW,in spite of the suppressed κL, signifies a decoupling of charge andlattice-driven heat transport. In the following, we present investiga-tions of the microstructure of these materials alongside local micro-scale probing of the Seebeck coefficient S. Finally, we show anddiscussexperimental results from extensive TE and magneto-transport mea-surements carried out in a broad range of temperatures and mag-netic fields.Structural modifications in FVABX compositesThe structural properties of the sintered samples were investigated viascanning and transmission electron microscopy (TEM), energy-dispersive X-ray spectroscopy (EDX), and X-ray diffraction (XRD).Fig. S4 shows BS-SEM images of Fe2V0.95Ta0.1Al0.95 sintered at 1373 Kwithout the addition of Bi0.9Sb0.1. Throughout the whole sample,nanoscale precipitates of a secondary Ta-rich phase are clearlynoticeable at the GBs. This is in agreement with the previously estab-lished low solubility limit of Ta, x =0.07 in Fe2V1−xTaxAl30. Apart fromthat, the microstructure displays a very homogeneous phase dis-tribution without any variations in the composition.Figure 2a shows a low-magnification image of the microstructureof the FVAB50 composite, with the best TE properties. A uniformdistribution of Bi1−xSbx along the GBs is obvious and confirmed bycompositional mapping using EDX (Fig. 2b) with an estimated volumefraction of around 5–7 vol. %. Additionally, we find that the segregationof Bi1−xSbx along the GBs goes hand in hand with two changes in themicrostructure: (i) strong suppression of nanoscale Ta-rich pre-cipitates at the GBs, (ii) diffuse brightness variations within the grains.Both these structural changes suggest an enhanced solubility limit ofheavy Ta atoms, when Bi1−xSbx is incorporated as a GB network duringthe liquid-phase sintering, contributing to a reduction of the latticethermal conductivity as shown later. This is confirmed by EDX linescans (Fig. 2d) across the grain, revealing periodic fluctuations in theTa and V concentration, and XRD, revealing an increase of the latticeparameter of the Heusler phase (Fig. 2e and inset of Fig. S12c) as largerand heavier Ta atoms are substituted. In Fig. 2f, g, we focus on Bi1−xSbx.Since Fe2VAl andBi1−xSbx are chemically and structurally distinct, thereexists a well-defined GB without apparent interdiffusion. We find thatBi1−xSbx, when embedded between Heusler grains, displays a peculiarladder-like nanostructure with arrays of stacking fault defects. More-over, detailed EDX analyses of different samples indicate that the Bi:Sbratio fluctuates and that the Sb concentration is above the nominal oneFig. 2 | Microstructure evolution in Fe2V0.95Ta0.1Al0.95 Heusler compoundsupon incorporating Bi1−xSbx. a Microstructure of FVAB50 composite, where themajority of GBs are filled with Bi-Sb. b EDX analyses reveal that Bi and Sb are foundat the GBs, while Fe, V, Al, and Ta are almost exclusively distributed within thegrains. c BS-SEM image of a Heusler grain with periodic contrast variations, sur-rounded by Bi-Sb. d EDX line scan along the Heusler grain shown in (c).e Comparison of normalized X-ray diffraction peaks of the (220) plane ofFe2V0.95Ta0.1Al0.95 and FVAB50 composite. f Bright-field TEM image of the GB withladder-likenanostructure arrays of stacking faults and (g) high-magnification imagenear stacking-faults. Insets in (e and g) show Heusler and Bi-Sb unit cells,respectively.Article https://doi.org/10.1038/s41467-025-57250-6Nature Communications |         (2025) 16:2976 3www.nature.com/naturecommunications(although remaining within the topological-insulating regime) in ourhigh-performance FVAB50 composite (Figs. S9 and S11).Thermoelectric propertiesThe pivotal role of understanding and investigating TE transportacross grain boundaries is increasingly recognized25,29,31,32. To draw aconnection between microstructure and electronic transport weemployed a transient potential Seebeckmicroprobe (TPSM), with localproperty investigations performed on the same rectangular area of thesample (Fig. 3a). Figure 3b shows a map of the locally determined Swith a spatial resolution of 3–5 microns. The results obtained are inexcellent agreement with structural investigations revealing a rich andcomplexmicrostructure consisting of Heusler grains and a Bi1−xSbxGBnetwork. Interestingly, TPSM measurements suggest that Bi1−xSbxexhibits a larger S as a secondaryphase compared to its bulk form. Thisis emphasized by looking at line scans across Heusler grains. The pla-teau in Fig. 3c refers to the value of Bi1−xSbx within the composite,which is significantly higher than its bulk value, especially consideringthat TPSMmeasurements typically underestimate S by at least 20–30%due to the cold finger effect. This enhancement, which exceeds thehighest S at 300K in the entire composition range of polycrystallineBi1−xSbx33, is surprising, given the near-complete immiscibility betweenBi1−xSbx and the Heusler phase.Figure 3d shows the distribution histogram of the measuredSeebeck coefficient. For the Heusler phase, S1 ≈ −140μVK−1, and forBi0.9Sb0.1, S2 ≈ −78μVK−1 would be expected. However, instead of twonormal distributions centered around those values, the observed dis-tribution appears much more merged with S being significantlyunder(over)estimated with respect to S1(S2). While the under-estimation is an artifact from the cold finger effect, inherent to theTPSM and basically all microprobemeasurements34, the enhanced S ofthe secondary phase indicates a beneficial interplay between the twocomponents and explains why the integral value of S remains large inthe composite (Fig. 4c), despite being short-circuited across the GBs.Wenote that a similar observation hasbeenmade already several yearsago by Mikami and Kobayashi in (Fe2VAl0.9Si0.1 + Bi) compositeswith zTmax =0:1135.In Fig. 4, we compare the temperature-dependent bulk TE prop-erties of our FVABX (X =0, 20, 50) composites over a broad tempera-ture range from 4 to 523K. Measurements were performed usingdifferent setups in various laboratories at the National Institute forMaterials Science (NIMS) in Japan and at TU Wien (TUW) in Austria.Additionally, extensive measurements have also been conducted on abulk sample of Bi0.9Sb0.1 synthesized during this study, and the datahave been included for comparison. The latter are in excellent agree-ment with those reported previously (see Fig. S13).Figure 4a displays the temperature-dependent thermal con-ductivity κ(T). At 200–300K, κ(T) increases due to bipolar thermaltransport, consistent with the S(T) curves. Most importantly, whenBi1−xSbx is incorporated as a secondary phase, κ(T) decreases sig-nificantly, which we attribute to the complex microstructural evolu-tion involvingmicroscale Bi1−xSbxGBswith an extremely large acousticmismatch (≈9THz) with respect to the Heusler matrix, periodic com-positionfluctuationswithin the grains, and an enhanced solubility limitof heavy Ta atoms. Moreover, κ(T) of the Bi1−xSbx GB network is likelyreduced as well compared to the bulk values owing to the ladder-likearrays of stacking fault defects (see Fig. 2f, g) and microscale compo-sition fluctuations (Fig. S9), likely inhibiting phonon-driven heattransport along the GBs36.FromFig. 4b, it is evident that, despite the significant reduction inκ(T), electronic transport remains excellent. The temperature-dependent resistivity ρ(T) flattens when Bi0.9Sb0.1 is incorporated,even resulting in a decrease of ρ(T) at elevated temperatures. Theflattening of the resistivity curves and the increased residual resistivityat low temperatures both imply a weakening of the electron-phononcoupling and enhanced disorder, aligning with the notion of anenhanced Ta solubility limit. Previous work shows that V/Ta substitu-tion results in a softening of the Heusler lattice and a reduction of theFig. 3 | Local investigation of electronic transport in (Fe2V0.95Ta0.1Al0.95 +Bi0.9Sb0.1) composites. a Microstructure, composition, and local transport prob-ing in the same area (orange square in Fig. 2a). b, c Transient potential Seebeckmicroprobe (TPSM)mapping of FVAB50 at room temperature. Ta-enriched regionsinside the Heusler grains and Bi1−xSbx at the GBs display a smaller S than theremaining part of the matrix. c TPSM line scans across distance marked in (b).d Histogram from TPSMmapping. Solid lines are normal distributions centeredaround S1 = −140μV−1 K−1 and S2 = −78μV−1 K−1, the bulk values for Fe2V0.95Ta0.1Al0.95and Bi0.9Sb0.1, respectively.Article https://doi.org/10.1038/s41467-025-57250-6Nature Communications |         (2025) 16:2976 4www.nature.com/naturecommunicationselectron-phonon deformation potential, decreasing ρ(T) at elevatedtemperatures, where acoustic phonon scattering dominates. More-over, V/Ta substitution can expand the band gap by pushing the V-egconduction band toward higher energies, enhancing the maximum ofthe Seebeck coefficient37.The temperature-dependent Seebeck coefficient S(T) only variesmoderately in the composites with Smax shifting to slightly lowertemperatures. As mentioned in the previous section, a simpleeffective-medium theory (EMT) with parallel conduction along the GBnetwork would result in a sizeable decrease of S(T). The fact that Sretains large values, is surprising and unexpected, calling for moreadvanced theoretical studies of TE transport in composite materials.Although it is well known that the TE properties of nanocompositescan deviate strongly from those of the individual materialcomponents38,39, deviations from the EMT in microscale compositesare much rarer.The above-listedmodifications result in an extremeboost of zT bymore than a factor of two (see Fig. 4d) up to a zTmax of almost 0.5 at295 K. Note that, like for almost all TE studies, error bars of ≈20%should be considered40, which were omitted for better visibility of thedata. This clearly exceeds the predictions of the EMT, which, asdemonstrated by Bergman and Levy in their seminal work41, states thatzT in composites needs to be always smaller than the largest zT of theindividual components, irrespective of the geometry. We also notethat this exceeds the largest room-temperature zT of the entire binaryBi-Sb system,with zTmax � 0:333. This implies that the TE properties ofthe individual components change dramatically in the composite orare subject to reciprocal action of both, allowing for a decoupling ofcharge and heat transport. To investigate the thermal stability of ourcomposites, transport measurements were conducted for variousthermal cycles (Fig. S22), which reveal excellent reproducibility and nodegradation of the properties at least up to 500K – the most relevanttemperature range for the potential application of these materials.Field-dependent magneto-transportTo further elucidate transport in the best-performing compositesample (FVAB50), we measured the Hall effect in a broad temperatureand magnetic field range, 4–400K and −9T to 9 T. These results aresummarized in Fig. 5. Thefield-dependentHall resistivityρxy, plotted inFig. 5a for various temperatures displays anextremely large anomalousHall effect, whicheven increaseswith rising temperature up to ≈300K,despite the absence of any sizeable magnetization in the sample. Onthe contrary, the sintered Heusler compound without Bi1−xSbx at theGBs exhibits a simple linear magnetic field dependence. While theobservation of a giant anomalous Hall effect in various topologicalmaterials is often ascribed to huge Berry curvatures, emerging fromthe respective topological band structure features42,43, we interpret thecomplex field-dependent curves in Fig. 5a as a two-channel conductionmechanism, where charge carriers can move across the sample eitherthrough topologically trivial bulk states of the Heusler grains orthrough topologically protected surface states of the Bi1−xSbx GBnetwork (see inset Fig. 5b). Figure 5b shows that the field-dependentbehavior of ρxy from −5 T to 5 T can be reasonably well described by asimple two-channel transport model (details of the modeling proce-dure and underlying theory is presented in the SI). The mobilitiesobtained for the two distinct transport channels are presented inFig. 5c alongside the values of pristine Fe2V0.95Ta0.1Al0.95 withouttopological-insulating GBs. The bulk values of Fe2V0.95Ta0.1Al0.95 are ofa bdcT (K)0 100 200 300 400 500 600-150-100-5000 100 200 300 400 500 60001002003004005006000 100 200 300 400 500 60002468101214T (K)0 100 200 300 400 500 600zT0.00.10.20.30.40.50.6κ(W m-1K-1)ρ( μΩcm)S(μV K-1)TPSM corr.TPSM (DLR)Bi-Sb phaseEM theorybulk Bi0.9Sb0.1FVAB50 compositeFVAB20 compositeFe2V0.95Ta0.1Al0.95TUWNIMSFig. 4 | Thermoelectric properties of (Fe2V0.95Ta0.1Al0.95 + Bi0.9Sb0.1) compo-sites. a Temperature-dependent thermal conductivity, b electrical resistivity,c Seebeck coefficient, and (d) dimensionless figure ofmerit zT of Fe2V0.95Ta0.1Al0.95composites with 20 and 50 vol.% Bi0.9Sb0.1 added before sintering (FVAB20,FVAB50) compared topristine Fe2V0.95Ta0.1Al0.95 andBi0.9Sb0.1. The error bars in (c)indicate the statistical variation of the Bi-Sbphase corresponding to the area shownin Fig. 3b. The red solid line in (d) represents a calculation based on effective-medium theory (EMT) for FVAB50, using a volume fraction of ≈6 vol.% Bi0.9Sb0.1,determined by EDX.Article https://doi.org/10.1038/s41467-025-57250-6Nature Communications |         (2025) 16:2976 5www.nature.com/naturecommunicationsthe order of 10 cm2 V−1 s−1. The mobility of the bulk channel, μ1,extracted from our transport modeling is comparable, especially atlow temperatures. The mobility of the Dirac-like surface states, μ2,associated with the Bi-Sb network, on the other hand, is several ordersof magnitude higher up to 3 × 105 cm2 V−1 s−1 and 2 × 104 cm2 V−1 s−1 forthe Dirac holes and electrons, respectively. As a consistency check, wecalculated the temperature-dependent zero-field resistivity from theobtained carrier mobilities μ1,2 and carrier densities n1,2 viaρxxð0,TÞ= ðen1μ1 + en2μ2Þ�1, which should match temperature-dependent measurements in Fig. 4b. As shown in Fig. S19, there isexcellent agreement across the entire temperature range, under-scoring the robustness and reliability of the fits.In summary, the field-dependent Hall effect reveals a significantcontribution to electronic transport from the Dirac-like surface statesof the Bi1−xSbx GB network, leading to a pronounced anomalous Halleffect, which can be explained by a two-channel transport model. Thisaligns with the colossal mobilities expected from such topologicallyrobust carriers and the higher surface-area-to-volume ratio in thecomposite.DiscussionConcluding, we demonstrated that incorporating Bi1−xSbx in Fe2VAlHeusler compounds can boost the zT compared to both individualmaterials. This is particularly surprising considering the near-completeimmiscibility of both components and their chemical distinctness,which should prevent sizeable interdiffusion and changes to the indi-vidual material properties. Decoupling charge and lattice-driven heattransport in such composites is an auspicious route toward high zT,even more so in systems where reducing grain size and alloyingstrongly compromise carrier mobility, although, a more profoundtheoretical understanding of charge and heat transport in compositesis required to optimally design and choose the best candidates.In this study, we achieved heavily reduced κL, and simultaneouslyhighμW. Toprovide a broad comparisonwith othermaterial classes forthe latter, we downloaded all available TE property data from theStarrydata2 open web database44, which, as of July 2024, contains TEdata from8961 different papers and 52,020different samples.We thencalculated μW for those samples, where both S(T) and ρ(T) are repor-ted. Fig. S23 shows that, near room temperature, μW of the bestcomposite sample from this work surpasses all other reported n-typesemiconductors.To further elevate the performance of Fe2VAl systems, broadscreening of secondary phases needs to be done; especially othertopological insulators like Bi2Se3 could be considered. Additionally,one has to think about strategies to increase and reliably tune thevolume fraction. Lastly, it is crucial to identify promising p-type com-pounds with competitive z. Since p-type Fe2VAl compounds inherentlyshow much smaller Seebeck coefficients, this can only be realized viaband engineering of the valence band electronic structure45,46. Onlythen can competitive Fe2VAl-based modules be realized, which couldsubstitute the long-reigning Bi2Te3 systems. The present study sug-gests that, byproperGBengineering, Fe2VAlHeusler alloysmay indeedbear the potential to rival state-of-the-art Bi2Te3 and Mg3Bi2 semi-conductors. High-performance modules entirely based upon Fe2VAlalloys could open a new era for near-ambient applications, as thesesystems excel in terms of cost-effectiveness, excellent recyclability,and simpler device structures. Moreover, they exhibit superiormechanical, thermal, and chemical long-term stability, factors that arebecoming increasingly recognized as essential assets for realizingwidespread thermoelectric technology.MethodsSynthesis of starting materials and compositesBulk elements of high purity (Fe 99.99%, V 99.93%, Ta 99.95%, Al99.999%, Bi 99.999%, Sb 99.999%) were stoichiometrically weighedand polycrystalline ingots of the starting materials (Fe2V0.95Ta0.1Al0.95and Bi0.9Sb0.1) were synthesized by high-frequency induction meltingunder Ar atmosphere. The as-cast ingots were manually crushed andground using a tungsten carbide pestle and mortar. The resultingpowders from the individual starting materials were then mixed invarious volume ratios. The volume percentage of Bi0.9Sb0.1 powderadded was adjusted and calculated based on the theoretical densitiesof the respective startingmaterials. After thepowderswere thoroughlymixed, the mixture was filled into a graphite die and sintered at atemperature of 1373 K, which is about 80% of the melting point of thefull-Heusler phase and about 800K higher than the melting point ofBi0.9Sb0.1. Consequently, excess liquid was expelled during the sin-tering process. Additionally, we observed that the liquid-phase sin-tering led to a significant decrease in the sintering temperature of theHeusler material by up to almost 200K when Bi0.9Sb0.1 powder wasadded as compared to when only Fe2V0.95Ta0.1Al0.95 powder was sin-tered. Nonetheless, to ensure consistent and comparable processingFig. 5 | Magneto-transport properties of FVAB50 composite. a Field-dependentHall resistivity of FVAB50 at different temperatures 4 ≤ T ≤ 400 K. Room-temperature data of Fe2V0.95Ta0.1Al0.95 are shown for comparison as black opencircles. b Two-channel transport modeling of field-dependent Hall resistivity. Solidlines are least squares fits. Inset shows a sketch of the Hall effect in FVABX com-posites. Highly mobile Dirac-like carriers along the Bi-Sb GB network have a muchlarger mean free path than the Heusler bulk electrons and are deflected mucheasier, even in small magnetic fields. c Temperature-dependent Hall mobility ofFVAB50 obtained by modeling the complex field-dependent behavior. Error barsindicate uncertainties in the fit results. The inset shows a sketch of the two trans-port channels (1) the bulk conduction electrons of the Heusler main phase and (2)the Dirac-like surface states of the Bi-Sb GB network.Article https://doi.org/10.1038/s41467-025-57250-6Nature Communications |         (2025) 16:2976 6www.nature.com/naturecommunicationsconditions for the samples studied in this work, all specimens weresinteredusing exactly the samesynthesis conditions, i.e., a compactionpressure of 50MPa, a maximum temperature of 1373 K, and a holdingtime of 15min. No additional heat treatment has been applied to thesamples after the sintering process.Structural characterizationThe microstructure and elemental composition of the sinteredsamples were investigated using scanning electron microscopy inboth secondary electron (SE) and backscattered electron (BSE)modes, complemented by energy-dispersive X-ray spectroscopy(EDX). These analyses were performed on an ultra-high-resolutionfield emission SEM (HRSEM SU8230, Hitachi, Japan), equipped withan X-MaxN EDS detector (Horiba, Japan). For HRSEM observations,the samples were mounted in electroconductive epoxy and polishedmeticulously. EDX analysis utilized an acceleration voltage of 25 kV,gathering 10 × 106 counts per EDX map and 1 × 106 counts for pointanalysis.To investigate the interface between the Fe2VAl matrix and theBi0.9Sb0.1 secondary phase at the nanoscale, the sample was preparedusing a conventional focused ion beam (FIB) technique. A thin sectionwas extracted from the targeted area, attached to an Omnigrid, andthinned to approximately 90 nm for electron transparency. Addition-ally, the FVAB50 sample was crushed into fine particles, dispersed inethanol, and deposited on a grid to investigate the sample surface.Transmission electron microscopy (TEM) bright-field and lattice ima-ges were acquired using a JEOL JEM-3100FEF (JEOL, Japan) microscopeoperating at 300 kV, whichwas also equippedwith an EDS detector fordetailed elemental mapping.The X-ray powder diffraction measurements were conducted atthe Institute of Solid State Physics, TU Wien, using an in-house dif-fractometer (AERIS by PANalytical). These measurements utilizedstandard Cu K-α radiation, with data collected in the Bragg-Brentanogeometry over the angular range 20∘ < 2θ < 100∘. Rietveld refinementson the obtained powder patterns were performed using the programPowderCell.High-temperature property measurementsThermal conductivity measurements at high temperatures were per-formed in N2 atmosphere directly on the sintered pellets, in thedirection parallel to the pressing (compaction) direction during sin-tering with a commercially available setup (LFA 467 by NETZSCH). Theinstrument makes use of a conventional laser flash method forthe diffusivity D and a differential scanning calorimeter for determin-ing the specific heat cp. The density of the sample dm was determinedvia Archimedes principle and the thermal conductivity was calculatedfrom κ =D cp dm.After performing high-temperature thermal conductivity mea-surements, the samples were cut into strips 2−3mm in widthand 8−10mm in length using a high-speed aluminum oxide cuttingwheel. The bar-shaped samples were then mounted in a commercialsetup (ZEM3 by ADVANCE RIKO) and the electrical resistivity andSeebeck coefficient were measured as a function of temperature. Forthe best sample, the measurement was repeated to confirm repro-ducible and stable results.Low-temperature property measurementsLow-temperature measurements provide valuable insights into lower-energy excited states and states near the Fermi energy. This is espe-cially significant for samples with narrow energy gaps, such as Fe2VAl-based full-Heusler and binary Bi1−xSbx systems, where the Seebeckcoefficient often peaks belowor near room temperature. Furthermore,when modeling temperature-dependent data (using a parabolic bandmodel for example), it is crucial that the experimental data span awidetemperature range. The thermoelectric characterization at lowtemperatures was carried out at TU Wien (Austria) on the same rec-tangular bar-shaped sample pieces that were used for the high-temperature measurements at NIMS (Japan).The temperature-dependent electrical resistivity was measuredin a home-built bath cryostat at TU Wien. The sample was contactedin a four-probe geometry with thin gold wires, using a spot-weldingdevice. The sample was then mounted on a sample puck using GEVarnish as an adhesive and directly inserted into the cryostat. Themeasurement was performed continuously every time the tempera-ture changed by 1 K.The temperature-dependent Seebeck coefficient was also mea-sured on the very same sample piece using a different home-builtsetup at TU Wien. Here, two chromel-constantan thermocouples arecontacted to both ends of the sample to pick up the temperaturedifference and voltages. Since it is difficult to solder directly on thesample surface, a bundle of thick copper wires was first spot-weldedonto the sample surface to which the thermocouples were then sol-dered. The high thermal conductivity of copper and the fact that thethermocouples are soldered in very close proximity to the samplesurface ensures that the cold finger effect can be minimized. Fur-thermore, two strain gauges with a resistance of ≈120Ω function asheaters and arefixed to the bottomof both sample ends viaGE varnish.The two heaters allow switching the temperature difference ("seesawheating"47) to cancel spurious voltage contributions.Themeasurementis carried out in anevacuated sample chamberwithHeexchange gas toensure thermal coupling to the cryogen.The thermal conductivity at low temperatures was measured bymaking use of a steady-state method using a home-built sample probewith a flow cryostat. Here, a heater is attached to the top surface of thesample employing a thermally conductive epoxy resin (STYCAST2850FT). Similar to the Seebeck coefficient measurements, two bun-dles of copper wires are first fixed to the sample to each of which athermocouple is then soldered. The bottom of the sample is mountedon a copper heat sink and the measurement is carried out in highvacuum ( ≈10−5 mbar).Modeling of field-dependent Hall resistivityIn a two-channel transportmodel for two types of charge carriers withcharge q1,2, carrier density n1,2, and carrier mobility μ1,2, the field-dependent longitudinal resistivity ρxx(B) and Hall resistivity ρxy(B) canbe expressed as48ρxxðBÞ=q1 n1 μ1 +q2 n2 μ2 + ðq1 n1 μ2 + q2 n2 μ1Þμ1 μ2 B2ðq1n1μ1 +q2n2μ2Þ2 + ðq1n1 +q2n2Þ2 μ21 μ22 B2 ð1ÞandρxyðBÞ=q1 n1 μ21 +q2 n2 μ22 + ðq1n1 +q2n2Þ2 μ21 μ22 B2ðq1n1μ1 +q2n2μ2Þ2 + ðq1n1 + q2n2Þ2 μ21 μ22 B2 B : ð2ÞFrom the above equations, it becomes evident that non-linearitiescan occur in the field-dependent Hall resistivity if, for instance, thereis a sizeable difference between μ1 and μ2. While previousmodels have employed four different fit parameters, i.e. therespective carrier densities and mobilities to model non-linear fielddependencies, Eguchi and Paschen previously suggested a novel,more robust scheme for analyzing field-dependent magneto-trans-port properties49. We utilized this framework tomodel the non-linearfield-dependent Hall effect of the FVAB50 composite. More detailsregarding the modeling procedure are presented in the Supplemen-tary Information.Data availabilityAll data supporting the findings of this study are available within thearticle and its Supplementary Information file.Article https://doi.org/10.1038/s41467-025-57250-6Nature Communications |         (2025) 16:2976 7www.nature.com/naturecommunicationsReferences1. Rowe, D. M. Thermoelectrics Handbook: Macro to Nano (CRCPress, 2018).2. Hendricks, T., Caillat, T. &Mori, T. Keynote reviewof latest advancesin thermoelectric generation materials, devices, and technologies2022. Energies 15, 7307 (2022).3. Zevalkink, A. et al. A practical field guide to thermoelectrics: fun-damentals, synthesis, and characterization. Appl. Phys. Rev. 5,021303 (2018).4. Snyder, G. J. et al. Weighted mobility. Adv. Mater. 32, 2001537(2020).5. Goldsmid, H. J. The electrical conductivity and thermoelectricpower of bismuth telluride. Proc. Phys. Soc. 71, 633–646(1958).6. Witting, I. T. et al. The thermoelectric properties of bismuth tell-uride. Adv. Electron. Mater. 5, 1800904 (2019).7. Mao, J. et al. High thermoelectric cooling performance of n-typeMg3Bi2-based materials. Science 365, 495–498 (2019).8. Liu, Z. et al. Demonstration of ultrahigh thermoelectric efficiency of~7.3% in Mg3Sb2/MgAgSb module for low-temperature energyharvesting. Joule 5, 1196–1208 (2021).9. Liu, Z. et al. Maximizing the performance of n-type Mg3Bi2 basedmaterials for room-temperature power generation and thermo-electric cooling. Nat. Commun. 13, 1120 (2022).10. Pecunia, V. et al. Roadmap on energy harvesting materials.J. Phys. Mater. 6, 042501 (2023).11. Anand, S. et al. Thermoelectric transport of semiconductor full-Heusler VFe2Al. J. Mater. Chem. C. 8, 10174–10184 (2020).12. Alleno, E. Review of the thermoelectric properties in nanos-tructured Fe2VAl. Metals 8, 864 (2018).13. Mikami, M., Kinemuchi, Y., Ozaki, K., Terazawa, Y. & Takeuchi, T.Thermoelectric properties of tungsten-substituted Heusler Fe2VAlalloy. J. Appl. Phys. 111, 093710 (2012).14. Renard, K. et al. Thermoelectric properties of the Heusler-type,Fe2VTaxAl1−x alloys. J. Appl. Phys. 115, 033707 (2014).15. Hinterleitner, B. et al. Stoichiometric and off-stoichiometric fullHeusler Fe2V1−xWxAl thermoelectric systems. Phys. Rev. B 102,075117 (2020).16. Furuta, Y., Kato, K., Miyawaki, T., Asano, H. & Takeuchi, T. Fe2VAl-based thermoelectric thin films prepared by a sputtering techni-que. J. Electron. Mater. 43, 2157–2164 (2014).17. Hiroi, S., Mikami, M., Kitahara, K. & Takeuchi, T. Thickness depen-dence of thermal conductivity and electron transport properties ofFe2VAl thin-films prepared by RF sputtering technique. Int. J.Nanotechnol. 13, 881–890 (2016).18. Bourgault, D. et al. Improvedpower factor in self-substituted Fe2VAlthermoelectric thin films prepared by co-sputtering. ACS Appl.Energy Mater. 6, 1526–1532 (2023).19. Mikami, M., Matsumoto, A. & Kobayashi, K. Synthesis and thermo-electric properties of microstructural Heusler Fe2VAl alloy. J. Alloy.Compd. 461, 423–426 (2008).20. Masuda, S., Tsuchiya, K., Qiang, J., Miyazaki, H. & Nishino, Y. Effectof high-pressure torsion on the microstructure and thermoelectricproperties of Fe2VAl-based compounds. J. Appl. Phys. 124,035106 (2018).21. Nishino, Y., Deguchi, S. & Mizutani, U. Thermal and transportproperties of the Heusler-type Fe2VAl1−xGex (0≤x≤0.20)alloys: Effect of doping on lattice thermal conductivity,electrical resistivity, and Seebeck coefficient. Phys. Rev. B74, 115115 (2006).22. Miyazaki, H., Tanaka, S., Ide, N., Soda, K. & Nishino, Y. Thermo-electric properties of Heusler-type off-stoichiometric Fe2V1+xAl1−xalloys. Mater. Res. Express 1, 015901 (2013).23. Takeuchi, T., Terazawa, Y., Furuta, Y., Yamamoto, A. & Mikami, M.Effect of heavy element substitution and off-stoichiometriccomposition on thermoelectric properties of Fe2VAl-based Heuslerphase. J. Electron. Mater. 42, 2084–2090 (2013).24. Fukuta, K., Tsuchiya, K., Miyazaki, H. & Nishino, Y. Improving ther-moelectric performance of Fe2VAl-based Heusler compounds viahigh-pressure torsion. Appl. Phys. A 128, 184 (2022).25. Wood,M., Kuo, J. J., Imasato, K. & Snyder, G. J. Improvement of low-temperature zT in a Mg3Sb2–Mg3Bi2 solid solution via Mg-vaporannealing. Adv. Mater. 31, 1902337 (2019).26. Luo, T. et al. Nb-mediated grain growth and grain-boundary engi-neering in Mg3Sb2-based thermoelectric materials. Adv. Funct.Mater. 31, 2100258 (2021).27. Qiu, Q. et al. Grain boundary scattering of charge transport inn-type (Hf,Zr)CoSb half-Heusler thermoelectric materials. Adv.Energy Mater. 9, 1803447 (2019).28. Bueno Villoro, R. et al. Grain boundary phases in NbFeSb half-heusler alloys: a new avenue to tune transport properties of ther-moelectric materials. Adv. Energy Mater. 13, 2204321 (2023).29. Mori, T. & Hara, T. Hybrid effect to possibly overcome the trade-offbetween Seebeck coefficient and electrical conductivity. Scr.Mater. 111, 44–48 (2016).30. Garmroudi, F. et al. Solubility limit and annealing effects on themicrostructure & thermoelectric properties of Fe2V1−xTaxAl1−ySiyHeusler compounds. Acta Mater. 212, 116867 (2021).31. Kuo, J. J. et al. Grain boundary dominated charge transport inMg3Sb2-based compounds.Energy Environ. Sci. 11, 429–434 (2018).32. Ghosh, S. et al. Towards energy filtering in Mg2X-based compo-sites: investigating local carrier concentration and band alignmentvia SEM/EDX and transient Seebeck microprobe analysis. Mater.Today Phys. 38, 101244 (2023).33. Doroshenko, A., Rogacheva, E., Drozdova, A., Martynova, K. &Men’shov, Y. V. Thermoelectric properties of polycrystallineBi1−xSbx solid solutions in the concentration range x = 0 − 0.25. J.Thermoelectr. 4, 23–36 (2016).34. Ziolkowski, P., Karpinski, G., Dasgupta, T. & Mueller, E. Probingthermopower on the microscale. Physica Status Solidi A 210,89–105 (2013).35. Mikami, M. & Kobayashi, K. Effect of Bi addition on microstructureand thermoelectric properties of Heusler Fe2VAl-sintered alloy. J.Alloy. Compd. 466, 530–534 (2008).36. Isotta, E. et al. Heat Transport at silicon grain boundaries. Adv.Funct. Mater. 34, 2405413 (2024).37. Garmroudi, F. et al. Boosting the thermoelectric performance ofFe2VAl-type Heusler compounds by band engineering. Phys. Rev. B103, 085202 (2021).38. Zhao, W. et al. Magnetoelectric interaction and transport beha-viours in magnetic nanocomposite thermoelectric materials. Nat.Nanotechnol. 12, 55–60 (2017).39. Zhao, W. et al. Superparamagnetic enhancement of thermoelectricperformance. Nature 549, 247–251 (2017).40. Borup, K. A. et al. Measuring thermoelectric transport properties ofmaterials. Energy Environ. Sci. 8, 423–435 (2015).41. Bergman, D. J. & Levy, O. Thermoelectric properties of a compositemedium. J. Appl. Phys. 70, 6821–6833 (1991).42. Nayak, A. K. et al. Large anomalous Hall effect driven by a non-vanishing Berry curvature in the noncolinear antiferromagnetMn3Ge. Sci. Adv. 2, e1501870 (2016).43. Manna, K. et al. From colossal to zero: controlling the anomalousHall effect in magnetic Heusler compounds via Berry curvaturedesign. Phys. Rev. X 8, 041045 (2018).44. Katsura, Y. et al. Data-driven analysis of electron relaxation timesin PbTe-type thermoelectric materials. Sci. Technol. Adv. Mater.20, 511–520 (2019).45. Nishino, Y., Kamizono, S., Miyazaki, H. & Kimura, K. Effects of off-stoichiometry and Ti doping on thermoelectric performance ofFe2VAl Heusler compound. AIP Adv. 9, 125003 (2019).Article https://doi.org/10.1038/s41467-025-57250-6Nature Communications |         (2025) 16:2976 8www.nature.com/naturecommunications46. Parzer, M. et al. High solubility of Al and enhanced thermoelectricperformancedue to resonant states in Fe2VAlx.Appl. Phys. Lett. 120,071901 (2022).47. Resel, R. et al. Thermopower measurements in magnetic fields upto 17 tesla using the toggled heating method. Rev. Sci. Instrum. 67,1970–1975 (1996).48. Ashcroft, N.W. &Mermin, N. D.Solid State Physics 1st edn (HarcourtCollege Publishers, 1976).49. Eguchi, G. & Paschen, S. Robust scheme for magnetotransportanalysis in topological insulators. Phys. Rev. B 99, 165128 (2019).50. Rogl, G. et al. (V, Nb)-doped half Heusler alloys based on {Ti, Zr, Hf}NiSn with high ZT. Acta Mater. 131, 336–348 (2017).51. Chauhan, N. S. et al. Vanadium-doping-induced resonant energylevels for the enhancement of thermoelectric performance in Hf-free ZrNiSn half-Heusler alloys. ACS Appl. Energy Mater. 1, 757–764(2018).52. Chauhan, N. S. et al. Compositional tailoring for realizing highthermoelectric performance in hafnium-free n-type ZrNiSn half-Heusler alloys. ACSAppl. Mater. interfaces 11, 47830–47836 (2019).53. Sankhla, A. et al.Mechanical alloying of optimizedMg2 (Si, Sn) solidsolutions: understanding phase evolution and tuning synthesisparameters for thermoelectric applications. ACS Appl. EnergyMater. 1, 531–542 (2018).54. Meng, X. et al. Grain boundary engineering for achieving highthermoelectric performance in n-type skutterudites. Adv. EnergyMater. 7, 1602582 (2017).55. Zhang, Q. et al. Deep defect level engineering: a strategy of opti-mizing the carrier concentration for high thermoelectric perfor-mance. Energy Environ. Sci. 11, 933–940 (2018).AcknowledgementsResearch in this paper was financially supported by the Japan Scienceand Technology Agency (JST) program MIRAI, JPMJMI19A1. Further-more, F.G. acknowledges financial support by the Lions Club Wien St.Stephan and J.d.B. acknowledges funding by the Deutsche For-schungsgemeinschaft (DFG, German Research Foundation), projectnumber 520487260 The authors also acknowledge the TU Wien Bib-liothek for financial support through its Open Access FundingProgramme.Author contributionsF.G., M.P., and A.R. conceived the idea for the study. F.G. designed thework, and, with help from I.S., C.B., and Y.H., synthesized the samplesand measured the thermoelectric properties. I.S., S.G., and H.D.N.investigated the micro- and nanostructure of the material via SEM andTEM techniques. S.G., P.Z., and G.O. performed TPSM measurements,and E.M. and J.d.B. analyzed and interpreted the data. C.B., S.S., G.R.,and E.S. assisted in the synthesis and experimental investigations of thesamples. P.R., N.K., E.M., and E.B. discussed and improved the contentsof the paper. J.d.B. and T.M. supervised the work and assisted in out-lining the initial draft of the manuscript. F.G. wrote the initial draft. Allauthors read, discussed, and edited the manuscript.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-025-57250-6.Correspondence and requests for materials should be addressed toFabian Garmroudi, Johannes de Boor or Takao Mori.Peer review information Nature Communications thanks Ajay Soni,Wenyu Zhao, and the other, anonymous, reviewer(s) for their contribu-tion to the peer review of this work. A peer review file is available.Reprints and permissions information is available athttp://www.nature.com/reprintsPublisher’s note Springer Nature remains neutral with regard to jur-isdictional claims in published maps and institutional affiliations.Open Access This article is licensed under a Creative CommonsAttribution 4.0 International License, which permits use, sharing,adaptation, distribution and reproduction in any medium or format, aslong as you give appropriate credit to the original author(s) and thesource, provide a link to the Creative Commons licence, and indicate ifchanges were made. The images or other third party material in thisarticle are included in the article’s Creative Commons licence, unlessindicated otherwise in a credit line to the material. If material is notincluded in the article’s Creative Commons licence and your intendeduse is not permitted by statutory regulation or exceeds the permitteduse, you will need to obtain permission directly from the copyrightholder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 2025Article https://doi.org/10.1038/s41467-025-57250-6Nature Communications |         (2025) 16:2976 9https://doi.org/10.1038/s41467-025-57250-6http://www.nature.com/reprintshttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/www.nature.com/naturecommunications Decoupled charge and heat transport in Fe2VAl composite thermoelectrics with topological-insulating grain boundary networks Results Decoupling charge and heat transport in Fe2VAl Structural modifications in FVABX composites Thermoelectric properties Field-dependent magneto-transport Discussion Methods Synthesis of starting materials and composites Structural characterization High-temperature property measurements Low-temperature property measurements Modeling of field-dependent Hall resistivity Data availability References Acknowledgements Author contributions Competing interests Additional information