# Fileset

[Advanced Materials---Active Diffusion Controlled Dual Stability in Thermoelectrics for Sustainable Heat.pdf](https://mdr.nims.go.jp/filesets/c89b2296-758e-4964-8a78-e873754bb6fd/download)

## Creator

[Longquan Wang](https://orcid.org/0009-0009-9910-9770), [Airan Li](https://orcid.org/0009-0004-7318-4821), Xinzhi Wu, Jiankang Li, [Takeo Ohsawa](https://orcid.org/0000-0001-7528-8940), [Takao Mori](https://orcid.org/0000-0003-2682-1846)

## Rights

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

## Other metadata

[Active Diffusion Controlled Dual Stability in Thermoelectrics for Sustainable Heat Harvesting](https://mdr.nims.go.jp/datasets/7107856a-af3d-404e-bc2d-1a3622b9f872)

## Fulltext

Active Diffusion Controlled Dual Stability in Thermoelectrics for Sustainable Heat HarvestingRESEARCH ARTICLEwww.advmat.deActive Diffusion Controlled Dual Stability in Thermoelectricsfor Sustainable Heat HarvestingLongquan Wang, Airan Li, Xinzhi Wu, Jiankang Li, Takeo Ohsawa, and Takao Mori*Thermoelectric technology offers a promising pathway toward globalsustainability by harvesting waste heat. However, long-term stability ishindered by inevitable elemental diffusion, degrading both the thermoelectricjunction and material properties, which prevents the realization of powergeneration applications. Here, dual and superior stability is achieved inhigh-performance Mg3(Bi,Sb)2, surpassing prior studies that focus on eitherjunction or material stability. By introducing an Mg layer at the junction,detrimental Mg diffusion is suppressed and compensate for Mg loss in thematerial, effectively stabilizing both junctions and materials for over 100 days.As a result, a thermoelectric module with 30-day-aged Mg3(Bi,Sb)2 is able tomaintain an outstanding power density of 0.45 W cm−2 and remarkableconversion efficiency of 8.6%, demonstrating unprecedented stability. Thesefindings provide new insights into thermoelectric junction engineering,shifting from interface optimization to comprehensive stabilization, advancingthe practical viability of thermoelectric energy harvesting for renewable andwaste heat applications.1. IntroductionThe transition from centralized fossil fuel systems to renewableand sustainable technologies aligns with the United Nations Sus-tainable Development Goals.[1,2] However, all thermodynamicprocesses inevitably produce heat loss due to the irreversibility ofL. Wang, A. Li, X. Wu, J. Li, T. MoriResearch Center for Materials Nanoarchitectonics (MANA)National Institute for Materials Science (NIMS)Namiki 1-1, Tsukuba 305-0044, JapanE-mail: MORI.Takao@nims.go.jpJ. Li, T.MoriGraduate School of Pure andAppliedSciencesUniversity of TsukubaTennodai 1-1-1, Tsukuba 305–8671, JapanT.OhsawaResearchCenter for Electronic andOpticalMaterialsNational Institute forMaterials Science (NIMS)Namiki 1-1, Tsukuba 305-0044, JapanThe ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/adma.202508270© 2025 The Author(s). Advanced Materials published by Wiley-VCHGmbH. This is an open access article under the terms of the CreativeCommons Attribution-NonCommercial License, which permits use,distribution and reproduction in any medium, provided the original workis properly cited and is not used for commercial purposes.DOI: 10.1002/adma.202508270the Carnot heat engine. Globally, an esti-mated 200 exajoules (EJ) of low-grade wasteheat are dissipated annually,[1] representingan untapped energy source. Converting thisenvironmental waste heat into usable en-ergy with eco-friendly and sustainable tech-nologies has become a critical focus in thepursuit of a sustainable future. Thermoelec-tric (TE) devices offer a promising solutionby converting thermal gradients into elec-tricity through the movement of charge car-riers driven by heat.[3,4] TE devices enabledirect energy conversion from waste heatto electricity with minimal maintenance re-quirements, supporting a sustainable en-ergy supply.[5,6] Despite significant progressin TEmaterials research, maintaining long-term high efficiency in TE devices re-mains a serious challenge for implemen-tation due to performance deterioration af-ter elemental diffusion during operation.TE devices rely on both TE materials andinterface materials (TEiMs) for efficient power generation.[7–9]Recent advancements have unveiled high-performance TE ma-terials, particularly near-room-temperature materials intendedto replace Bi2Te3, such as Mg3(Bi,Sb)2,[10–21] MgAgSb,[22–25] andAg2Q (Q = Te, Se, S).[26–29] At the same time, designing reli-able TEiMs is equally crucial, as they minimize electrical andthermal resistance at interfaces. Effective screening strategies,such as phase diagram calculations,[9,30,31] alloying,[32,33] inter-facial reaction and diffusion criteria,[34,35] and high-throughputmethods,[36,37] have identified TEiMs with low contact resistiv-ity (𝜌c). Despite these advancements, elemental diffusion at in-terfaces remains a persistent challenge, progressively degradingboth the TE material and its interface. Since diffusion is fun-damentally driven by thermodynamic equilibrium, it cannot becompletely suppressed, making long-term stability a critical yetunresolved issue.Traditional approaches primarily focus on minimizing 𝜌cto improve TE junction performance, assuming that stabil-ity issues arise solely from interfacial degradation. However,even junctions with initially low 𝜌c suffer from diffusion-driven degradation,[38] particularly in composition-sensitive TEmaterials,[39] where even minor elemental migration can dras-tically alter the transport properties of materials. For instance,in Bi2Te3-based materials, a minor shift in the Te mole frac-tion from 0.5997 to 0.6002 can induce a transition from p-typeto n-type conduction due to native defect evolution.[40] Simi-larly, in PbTe, interfacial diffusion perturbs the delicate balanceAdv. Mater. 2025, 37, 2508270 2508270 (1 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbHhttp://www.advmat.demailto:MORI.Takao@nims.go.jphttps://doi.org/10.1002/adma.202508270http://creativecommons.org/licenses/by-nc/4.0/http://creativecommons.org/licenses/by-nc/4.0/http://crossmark.crossref.org/dialog/?doi=10.1002%2Fadma.202508270&domain=pdf&date_stamp=2025-06-26www.advancedsciencenews.com www.advmat.debetween Pb-rich and Te-rich states,[41] affecting chemical poten-tial and doping efficiency. The issue is even more pronounced inMg3(Bi,Sb)2-based materials, where the high chemical reactivityand volatility of Mg exacerbate instability. In n-type Mg3(Bi,Sb)2,excess Mg modifies the chemical potential and suppresses Mgvacancy formation,[10,42] enabling n-type transport. However, un-controlled Mg migration at the interface can rapidly degrade ma-terial properties or even trigger a conduction-type reversal fromn-type to p-type, fundamentally altering device performance.N-type Mg3(Bi,Sb)2 has emerged as a leading candidate forlow-temperature thermal energy harvesting due to its high TEperformance, earth-abundant composition, and excellent me-chanical toughness.[43–45] Despite achieving high dimensionlessfigure-of-merit (zT) values of ≈1.0 at room temperature and ≈2.0at elevated temperatures,[18,46–49] practical application is severelylimited by chemical instability and thermal instability.[50–52] Con-tinuousMg reaction and evaporation leads to performance degra-dation, a challenge further exacerbated at TE junctions, where in-terfacial diffusion accelerates instability beyond that observed inbulk materials. As a result, achieving sustainable TE power gen-eration requires a holistic approach that stabilizes both the mate-rials and its interface—an issue that remains largely unsolved.Tomitigate these limitations, variousmetals such as Fe,[16,53–55]Ni,[13,56] and Nb,[36] were explored as TEiMs in Mg3(Bi,Sb)2 junc-tions. However, a sharp increase in 𝜌c after aging indicates sig-nificant interfacial diffusion or reaction, primarily driven bythe highly reactive nature of Mg. Even slight Mg migration atthe interface can degrade material performance.[42] To counter-act these effects, alloy-based TEiMs have been developed, in-cluding 304 stainless steel (304SS),[57] Cu7Ni3,[58] Fe7Mg2Cr,[32]Mg4.3Sb3Ni,[59] and FeCrTiMnMg,32 many of which incorporateMg to mitigate Mg loss from the material matrix. However, con-ventional TEiMs still exhibit Mg concentration gradients at in-terfaces, sustaining a chemical potential imbalance that perpet-uates diffusion-driven degradation. Material deterioration frominterfacial diffusion has been underappreciated, yet it poses asignificant barrier to achieving sustainable TE power genera-tion in Mg3(Bi,Sb)2 systems. The rapid degradation of mate-rial properties—exemplified by a shift from n-type to p-typeconductivity when Fe is used as a contact layer—[60]highlightsthe urgent need for dual stability in both materials andinterfaces.To address the long-standing stability challenges inMg3(Bi,Sb)2, we introduce anMg interlayer betweenMg3(Bi,Sb)2and the metallization layer to actively control the Mg diffusion.This interlayer effectively prevents Mg loss into the contactlayer while compensating for Mg depletion at grain boundaries,ensuring dual stability in both 𝜌c and material performance.As a result, the junction maintains holistic stability for over100 days at 573K. Even after 30 days of aging, TE modulesincorporating this dual-stable Mg3(Bi,Sb)2 junction achieve aremarkable conversion efficiency of 8.6% and a power densityof 0.45 W cm−2 under a relatively small 294 K temperaturegradient. These results highlight the promising potential of adual-stable Mg3(Bi,Sb)2 junction for sustainable low-grade heatharvesting. Furthermore, this strategy can be extended to otherTEmaterials sensitive to chemical potential variations, providinga generalizable solution for enhancing the performance andlongevity of TE devices.2. Results2.1. Achievement of Dual Stability in Mg3(Bi,Sb)2 JunctionsConventional TEiMs often exhibit Mg concentration gradients atinterfaces (Figure 1a), which drive diffusion due to differencesin chemical potential. This diffusion undermines the stability ofthe interfaces and leads to Mg loss, significantly impacting theTE properties of materials. In some cases, such as with Fe as theTEiM, this diffusion can even cause a transition in conductiontype from n- to p-type in Mg3(Bi,Sb)2 junctions.[60] In contrast,inserting an Mg interlayer creates a concentration barrier that ef-fectively suppresses Mg diffusion into the TEiM, while also com-pensating for Mg loss in the material. This dual mechanism sta-bilizes both the junction interface and the TE material.To verify this effect, we fabricated Mg/Mg3(Bi,Sb)2 junctionsand analyzed their thermal stability over extended periods. Lin-ear resistance scanning across the junction after aging at 573 Kreveals a stable resistance trend, indicating holistic stability ofthe junction (Figure 1b). The unchanged slope of the linear re-sistance curve in the material region confirms stable transportproperties. Evaluating stability at 573 K is directly relevant to real-world operating conditions for low-grade waste heat harvesting.In addition, the negligible resistance jump at the interface sug-gests a consistently low 𝜌c, which remains unaffected by aging.In contrast, conventional TEiMs typically exhibit a decline in 𝜎over time. This deterioration is primarily caused by Mg diffu-sion from the material to the contact layer, with the concentra-tion gradient at the interface accelerating this process. Notably,our Mg diffusion barrier is the first TEiM to maintain undete-riorated material performance in Mg3(Bi,Sb)2 junctions, achiev-ing superior stability compared to other TEiMs, and even outper-forming coating-based stabilization methods (Figure 1c).[33] Thisenhanced stability can be attributed to the effective preventionof Mg loss and the compensatory role of the interlayer, as willbe discussed later. Consequently, the Mg/Mg3(Bi,Sb)2 junctionmaintains stable material properties for over 50 days at 573 K,outperforming other junctions.Achieving a low 𝜌c is also critical for optimizing TEiMs, as high𝜌c impairs device performance. The power density and conver-sion efficiency of TE devices are significantly impacted by 𝜌c, asdescribed by the equation:[8]ZTD = zTM × L∕(L + 2𝜌c × 𝜎) (1)where L, 𝜌c, 𝜎, zTM, ZTD are the length of TE legs, contact resis-tivity, electrical conductivity, and the zT of materials and devices,respectively. For instance, to limit a 5% loss in ZTD/zTM for a de-vice with L= 2mmand 𝜎 = 105 Sm−1, 𝜌c must remain below 5 μΩcm2. As shown in Figure 1d, previously reported junctions exhibita significant increase in 𝜌c over time,[32,33,36,57,60,61] primarily dueto elemental diffusion at the interfaces. This rapid increase in 𝜌ccompromises the efficiency of the TE devices during operation.In contrast, the Mg/Mg3(Bi,Sb)2 junction demonstrated excep-tional stability, maintaining a low 𝜌c (<5 μΩ·cm2) over 30 days ofaging. The Mg interlayer effectively blocks elemental migration,preserves interface integrity, and ensures reliable long-term per-formance. Overall, the developed Mg/Mg3(Bi,Sb)2 junction ex-hibits distinct advantages over conventional designs, achievingAdv. Mater. 2025, 37, 2508270 2508270 (2 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.dewww.advancedsciencenews.com www.advmat.deFigure 1. Achievement of overall stability in TE junctions. a) Comparison between conventional TEiMs and Mg interlayer, highlighting the distinctMg concentration gradients at the interface. b) Evolution of linear resistance scanning across the Mg/Mg3(Bi,Sb)2 junction after aging at 573 K. c)Comparison of the 𝜎d/𝜎0 for materials in TE junctions,[33,60] where 𝜎0 is initial 𝜎 and 𝜎d is the 𝜎 after aging. d) Comparison of the evolution of 𝜌c overtime.[32,33,36,57,60,61]dual stability in both 𝜌c andmaterial performance by actively con-trolling the Mg diffusion at the interface.2.2. Superior Stability of TE Materials in TE JunctionsIn Mg3(Bi,Sb)2 materials, the potential diffusion and loss of Mghinder performance stability by introducing Mg vacancies. Thehigh saturated vapor pressure and volatility of Mg lead to sig-nificant diffusion during the sintering process of the junction,causing disrupted composition uniformity in Mg3(Bi,Sb)2. Thisphenomenon is evident from the poor linearity in the contactresistance curves observed in junctions with commonly usedTEiMs such as Fe andNi (Figure S1, Supporting Information).[60]Mg loss also introduces strong compositional fluctuations inMg3(Bi,Sb)2,[62] hindering the achievement of high TE perfor-mance. In contrast, the Mg/Mg3(Bi,Sb)2 joint demonstrates per-fect linearity in the contact resistance curve (Figure S2, Support-ing Information), indicating suppressed Mg loss during the sin-tering process and preserved composition homogeneity. More-over, continuous Mg diffusion occurs in conventional junctionswith a concentration gradient during the aging process, as ev-idenced by the progressively changing slopes of contact resis-tance curves. These changes indicate deteriorating 𝜎, as ob-served in the Ni/Mg3(Bi,Sb)2 and 304SS/Mg3(Bi,Sb)2 junctions(Figure 2a). In contrast, the Mg/Mg3(Bi,Sb)2 junction exhibitsAdv. Mater. 2025, 37, 2508270 2508270 (3 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.dewww.advancedsciencenews.com www.advmat.deFigure 2. Dual stability of material properties and 𝜌c. a) Evolution of linear resistance scanning across junctions after aging at 573 K. Temperature-dependent b) 𝜎, c) Seebeck coefficient of materials in different junctions. d) Evolution of 𝜌c and 𝜎 of the TE material in the Mg/Mg3(Bi,Sb)2 junctionover aging time at 573 K. e) Room-temperature 𝜌c of the Mg/Mg3(Bi,Sb)2 interface. f) UPS spectrum of Mg3(Bi,Sb)2 and Mg. g) Schematic illustrationof carrier transport at the ohmic contact interface.steady linearity in the contact resistance curves (Figure 1b), sug-gesting excellent thermal stability. Previous studies have demon-strated that dense Fe foil can effectively mitigate Mg diffusionfrom the material matrix to the interfacial layer.[60] However,the gradual increase in contact resistance after aging indicatesthat Mg diffusion still occurs due to the concentration gradient(Figure S3, Supporting Information).[63,64] In comparison, the de-veloped Mg/Mg3(Bi,Sb)2 junction displays superior thermal sta-bility relative to other TEiMs used in Mg3(Bi,Sb)2, including Ni,304 stainless steel (304SS), and Fe foil.Furthermore, the electrical transport properties of the agedTE junctions were investigated to evaluate material stability.The 𝜎 of materials in Ni/Mg3(Bi,Sb)2 junctions significantly de-creases across the entire temperature range compared to thepristine sample, with this trend intensifying as aging time in-creases (Figure 2b). Specifically, the room-temperature 𝜎 of theNi/Mg3(Bi,Sb)2 junction drops to 5.1 × 104 S m−1, a 43% reduc-tion compared to the pristine sample (9× 104 Sm−1). This declineis related to Mg diffusion from the material into the interfacelayer. However, after incorporating a Mg diffusion barrier, the𝜎 remains stable over extended aging periods and even slightlyexceeds that of the pristine sample. This stability suggests thatMg diffusion to the interface layers has been effectively blockedand that potential Mg compensation from theMg layer optimizesthe material’s electrical transport properties. The slight initial in-crease in 𝜎 observed during early-stage aging is likely due to im-proved compositional homogeneity and reduced grain-boundarypotential (Figure 1c).[62]A similar trend is observed in the Seebeck coefficient, asshown in Figure 2c. The sharply increased Seebeck coefficientin the Ni/Mg3(Bi,Sb)2 junctions indicates reduced carrier con-centration, driven by Mg loss. According to phase boundarymapping,[42] excess Mg is critical for maintaining n-type behav-ior in Mg3(Bi,Sb)2 and significantly impacts carrier concentra-tion. The observed material deterioration has a much strongerimpact on TE device performance than 𝜌c. For example, evenwhen 𝜌c increases to 20 μΩ cm2 in junctions with dimensionsof 3.8 × 3.8 × 6 mm3, the internal resistance contribution fromAdv. Mater. 2025, 37, 2508270 2508270 (4 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.dewww.advancedsciencenews.com www.advmat.detwo joints amounts for only ≈6.4% of the total, far lower thanthe performance loss caused by material deterioration (over 40%in Ni/Mg3(Bi,Sb)2 junctions). After aging at 573 K for 30 days,all tested junctions (Ni, 304SS, and Fe foil) show reductions in 𝜎of over 10%, except for the Mg/Mg3(Bi,Sb)2 junction (Figure 1c).To ensure the reliability of these results, we cross-validated the𝜎 values obtained from linear resistance scanning with thosemeasured using a standard four-probe method on a ZEM-3 sys-tem (Figure S4, Supporting Information), confirming consis-tency within 5% error.2.3. Simultaneously High Stability of Contact in Mg/Mg3(Bi,Sb)2JunctionsThe 𝜌c in a metal-semiconductor contact depends on the bar-rier height (ϕB) and carrier concentration (n), described by therelationship: 𝜌c ∼ exp(qϕB/n1/2),[8] where q is the elementarycharge. In other junctions, Mg diffusion reduces carrier concen-tration, leading to an increase in 𝜌c over time (Figure 1d). Conse-quently, blocking Mg diffusion from the material into the inter-face layers is crucial, not only for maintaining the thermal sta-bility of materials but also for ensuring a low 𝜌c. Remarkably,the Mg/Mg3(Bi,Sb)2 junction maintained an ultra-low 𝜌c (<5 μΩcm2) and stable material performance, including undeteriorated𝜎, over 100 days of aging at 573 K (Figure 2d). This underscoresthe critical importance of controlling Mg diffusion when fabri-cating reliable TE junctions. The aging duration of 100 days at573 K is significantly longer than most reported studies on TEjunction stability, and this stability is expected to extend beyond100 days, as suggested by the absence of degradation in electricalperformance and the overall robustness of the interface design.In our Mg/Mg3(Bi,Sb)2 junctions, an ultra-low 𝜌c of 1.71 μΩcm2 was achieved (Figure 2e), which is the lowest value reportedto date (Figure S5, Supporting Information). This result repre-sents the simplicity of module fabrication using metal contactsinstead of alloy contacts. To explore the origin of the low 𝜌c,the barrier height related to the work function (ϕ) of the metal-semiconductor contact was investigated.[65,66] For a contact be-tween an n-type semiconductor and a metal, achieving an Ohmiccontact is essential to minimize 𝜌c.[35] In such cases, the workfunction of the semiconductor (ϕs) should be greater than thatof the metal (ϕm). The work functions were calculated using theequation:𝜙 = h𝜈 −(Ecutoff − EFermi)(2)where h, 𝜈, Ecutoff and EFermi are Plank constant, the frequencyof the excited photons, the energy of the cut-off edge, and theenergy of the Fermi edge, respectively. Ultraviolet photoelectronspectroscopy (UPS)measurements were conducted to determineEcutoff, with the photon energy (h𝜈) set to 21.22 eV. The measuredEcutoff values for Mg3(Bi,Sb)2 and Mg were 18.05 and 18.16 eV,respectively (Figure 2f). Accordingly, the calculated work func-tions were 3.17 eV for Mg3(Bi,Sb)2 and 3.06 eV for Mg. Since ϕs> ϕm, an Ohmic contact is expected to form at the interface of thejunction. This expectation is further supported by the linear I–Vcharacteristics, with no rectification observed even after 30 daysof aging (Figure S6, Supporting Information).At the Mg/Mg3(Bi,Sb)2 interface, electrons transfer from theMg layer into Mg3(Bi,Sb)2 due to Fermi level alignment,[65,66]reaching equilibrium through charge redistribution. This chargetransfer induces an electron accumulation region on theMg3(Bi,Sb)2 side, increasing the local electron concentrationand enhancing conductivity. The favorable band alignment atthe Mg/Mg3(Bi,Sb)2 junction ensures barrier-free electron trans-port, enabling an Ohmic contact (Figure 2g).[8,65] These char-acteristics collectively account for the observed low 𝜌c in theMg/Mg3(Bi,Sb)2 junction. However, despite the formation of anOhmic contact, a small yet nonzero 𝜌c is still observed. This resid-ual 𝜌c likely originates from interface roughness, potential inter-facial impurities (e.g., thin oxide layers), and localized variationsin chemical composition, which may introduce minor scatter-ing effects. Nevertheless, the measured 𝜌c remains significantlylower than that of conventional TE junctions, confirming the su-perior electrical transport of the Mg/Mg3(Bi,Sb)2 interface. Thisultra-low 𝜌c remains stable, owing to the reliable stability of boththe contact layer and the TE material properties.2.4. Blocking Mg Diffusion from Materials to the Contact Layerand Mg CompensationTo uncover the origins of the remarkable stability observed inboth 𝜌c and material properties, we performed scanning electronmicroscope (SEM) imaging and energy-dispersive X-ray spec-troscopy (EDS) analysis. In the Ni/Mg3(Bi,Sb)2 junction, a promi-nent diffusion layer (≈10 μm thick) was observed after sintering(Figure 3a), primarily composed of Mg and Ni (Figure S7, Sup-porting Information). Moreover, significant compositional fluc-tuations, including Bi-rich phases within the materials, were de-tected. These compositional fluctuations are associated with Mgloss, which compromises the electrical transport properties ofMg3(Bi,Sb)2.[62,67] As hypothesized, the diffusion layer expandssignificantly after aging, increasing in thickness from≈10 to≈30μm (Figure 3b), further degrading the material’s performance.This growth was accompanied by the formation of Mg-Ni alloysnear the diffusion region, destabilizing the interface (Figure S8,Supporting Information). Evenwhen the TEiMwas replacedwithan alloy such as 304 stainless steel, Mg diffusion from the mate-rial to contact layer persisted (Figure 3c). The mechanism under-lying this material deterioration is illustrated in Figure 3d. Theconcentration gradient drives Mg diffusion, introducing abun-dant Mg vacancies within Mg3(Bi,Sb)2, which reduces carrierconcentration and 𝜎. Simultaneously, the unstable interface anddecreased carrier concentration contribute to increased 𝜌c afteraging.Remarkably, the Ni/Mg/Mg3(Bi,Sb)2 junction exhibits a well-defined interface structure after sintering, with the Mg layer (≈5μm thick) effectively blocking Mg element from the material intothe metallization layer (Figure 3e). Moreover, Ni from the metal-lization layer diffuses into the Mg layer, forming Mg2Ni, as con-firmed by the phase diagram.[68] This process protects the Mglayer from oxidation and strengthens the bonding at the Ni/Mginterface. EDS analysis further reveals that no matrix elementsdiffuse into the Ni layer, which remains composed entirely ofNi (Figures S9 and S10, Supporting Information). Notably, thisrobust sandwich structure remains intact even after extendedAdv. Mater. 2025, 37, 2508270 2508270 (5 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.dewww.advancedsciencenews.com www.advmat.deFigure 3. BlockingMg diffusion and enablingMg compensation. SEM images, EDSmapping, and line scanning results of different junctions with TEiMs:a) Ni after sintering, b) Ni after aging for 30 days, and c) 304SS after aging for 30 days. d) Schematic illustration of the interface structure following Mgdiffusion in conventional junctions. SEM images, EDS mapping, and line scanning results of Mg/Mg3(Bi,Sb)2 junctions: e) as-sintered (0 days), f) after10 days of aging, and g) after 30 days of aging. h) Comparison of the composition profile in the Mg/Mg3(Bi,Sb)2 junction after aging. i) Schematic ofMg diffusion blocking and potential Mg compensation mechanisms at the interface. j) Contact resistance curve of the junction in (i) after aging.aging (Figure 3f,g), demonstrating excellent thermomechanicalstability. Over time, Ni continues to diffuse into the Mg layer,forming additional Mg2Ni alloys and creating a metallurgicallybonded contact that further enhances interfacial strength (FigureS11, Supporting Information). Additionally, a ≈1 μm reactionlayer forms at the interface after sintering (Figure 3h), and Mgis known to form strong bonds with Mg3(Bi,Sb)2,[32] ensuringhigh shear strength in the Mg/Mg3(Bi,Sb)2 junction. Even after30 days of aging at 573 K, no apparent expansion of this reactionlayer is observed, further confirming its thermomechanical sta-bility. Importantly, the Mg layer disrupts the concentration gradi-ent drivingMg diffusion into the contact layers while consistentlysupplyingMg to the TEmaterial. This fully invertedMg diffusiontrend in our Mg/Mg3(Bi,Sb)2 junction contributes to its unprece-dented thermal stability compared to previous studies. However,compared to the pronounced Mg back-diffusion into the TE ma-terial, Mg diffusion into the Ni layer is minimal, owing to thedense Ni foil (Figure S11, Supporting Information).Microstructural characterization sheds light on the mecha-nisms driving the exceptional thermal stability of materials inthe Mg/Mg3(Bi,Sb)2 junction. A direct comparison between theMg/Mg3(Bi,Sb)2 joint and material without a contact layer high-lights two key effects of the Mg layer: inhibition of Mg diffusionand Mg compensation. As shown in Figure 3i, when no con-tact layer is present, the upper side of the junction suffers fromsignificant Mg loss during long-term operation at elevated tem-peratures. This Mg loss introduces Mg vacancies and hole car-riers, adversely impacting TE performance. Thermodynamically,Mg vacancies are more likely to form at grain boundaries thanwithin the bulk phase due to the higher Gibbs free energy at grainboundaries. This grain boundary diffusion mechanism, whichhas been previously validated,[51,69] contributes to Mg deficien-cies at grain boundaries.[70] These deficiencies further propagateinto the grains, creating a concentration gradient of Mg from thebulk to the grain boundaries and interfaces, accelerating the Mgloss process.In contrast, the insertion of an Mg layer establishes an Mg-rich environment at the interface, which is thermodynamicallyfavorable for blocking Mg diffusion pathways. Additionally, Mgcompensation from the interface to the grain boundary creates aMg-enriched grain boundary. This Mg-rich grain boundary effec-tively suppresses further Mg diffusion from the material bulk tothe grain boundary. This Mg compensation is expected to prop-agate throughout the material via grain boundary diffusion, off-setting surface-evaporation-induced Mg loss and thereby main-taining stable 𝜎 across the entire junction (Figure S2, Support-ing Information). As a result, grain-boundary potential barriersare significantly reduced, thereby enhancing electron transportin the material. This is supported by the slightly improved 𝜎 ob-served in the Mg/Mg3(Bi,Sb)2 junction (Figure 2b). Therefore,this Mg compensation effect not only enhances grain-boundarytransport by reducing potential barriers but also maintains car-rier concentration and suppresses increases in 𝜌c during aging.Long-term thermal stability tests at 573 K demonstrate the ef-fectiveness of the Mg layer. Initially, the junction exhibits perfectlinearity in the contact resistance curve and high 𝜎 due to theMg layer’s stabilization effects during sintering (Figure S12, Sup-porting Information). After sintering, the top contact layer wasAdv. Mater. 2025, 37, 2508270 2508270 (6 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.dewww.advancedsciencenews.com www.advmat.deFigure 4. Superior performance of TE module based on Mg/Mg3(Bi,Sb)2 junctions. Comparison of a) Rin and b) ɷmax among modules fabricated withdifferent junctions. The inset in (a) illustrates the two-pair module design, while the inset in (b) shows optical image of the Mini-PEM system used forTE module performance evaluation. c) 𝜂 as a function of I under different ΔT. d) Comparison of 𝜂max for modules with different junctions. e) ɷmax andf) 𝜂max of the two-pair module in this work compared with previously reported results from the literature.[9,24,33,36,49,53,71,72]removed by polishing for aging tests. Over time, Mg loss reduces𝜎, particularly on the top side, where it declines by ≈29% to ≈6 ×104 S m−1 after 50 days (Figure 3j). In contrast, the bottom sideretains excellent materials stability, maintaining 𝜎 at≈8.4 × 104 Sm−1. This highlights the effectiveness of theMg layer in blockingdiffusion and compensating Mg at grain boundaries, ensuringstable chemical potential and performance in materials. This ap-proach ensures unparalleled stability in both material propertiesand contact, representing a significant step forward in the devel-opment of long-lasting Mg3(Bi,Sb)2-based TE systems. While theMg interlayer is primarily responsible for suppressing diffusionand stabilizing the junction, the addition of a Ni layer enhancessolderability and supports practical device integration.2.5. Sustainable Power Generation in Modules usingMg/Mg3(Bi,Sb)2 JunctionsTo evaluate the potential of Mg/Mg3(Bi,Sb)2 junctions for long-term power generation, we fabricated two-pair TE modules, asillustrated in the inset of Figure 4a. The p-type MgAgSb was se-lected for its excellent near-room-temperature TE performance,with the TE properties of Mg3(Bi,Sb)2 and MgAgSb presentedin Figures S13 and S14 (Supporting Information). To minimizethe influence of the p-type junctions, Sb-based alloy was usedas the contact layer for p-type MgAgSb due to its low 𝜌c (FigureS15, Supporting Information).[38] Multiple TEmodules were fab-ricated by using aged n-type junctions (30 days at 573 K), includ-ing conventional TEiMs such as Ni and 304SS for direct compar-ison. Detailed data on output voltage (V), output power (P), heatflow from the cold side (Qc), and conversion efficiency (𝜂) are pre-sented in Figures S16–S18 (Supporting Information). The linear-ity of the current (I) versus V relationship enabled determinationof the internal resistance (Rin) from the slope. The module withNi as the contact layer exhibited an ultrahigh Rin, which cannotbe solely attributed to the increased 𝜌c (Figure 4a). A significantcontribution to the increased Rin came from material degrada-tion, as discussed earlier. Using alloy-based contact layers, suchas 304SS, partially delayed material deterioration, resulting in re-duced Rin compared to Ni. However, the Rin remained signifi-cantly higher than in themodule withMg/Mg3(Bi,Sb)2 junctions,consistent with prior observations. Our developed active diffu-sion control, employing an Mg layer, ensures the long-term ther-mal stability of both the 𝜌c andmaterial properties. This results inamarkedly lowerRin inmodules withMg/Mg3(Bi,Sb)2 junctions,where Rin = Rcontacts + Rmaterials. This improvement highlights thecritical role of the Mg layer in achieving enhanced performanceand durability in TE modules. Benefiting from the significantlylower Rin, the module with Mg/Mg3(Bi,Sb)2 junctions demon-strated a noticeably higher P. The relationship between Rin andP follows P = VocI − RinI2, where Voc is the open-circuit voltageof the module. A peak power density (ɷmax) of 0.45 W cm−2 wasachieved under a temperature gradient (ΔT) of 294 K, which isslightly lower than the theoretical prediction (Figure 4b). In con-trast, modules with Ni or 304SS as TEiMs exhibited only 65% and80% of the ɷmax, respectively, under the same ΔT.Remarkably, the module utilizing Mg/Mg3(Bi,Sb)2 junctionsachieved a maximum conversion efficiency (𝜂max) of 8.6% underAdv. Mater. 2025, 37, 2508270 2508270 (7 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.dewww.advancedsciencenews.com www.advmat.deΔT = 294 K, even after 30 days of aging (Figure 4c). Notably, noperformance deterioration was observed compared to modulesfabricated with unaged junctions (Figure S19, Supporting Infor-mation). In fact, a slight increase in efficiencywas observed, likelydue to enhanced electrical transport fromMg compensation andminor variations in the performance of the p-type junctions. Thesuperiority of the Mg/Mg3(Bi,Sb)2 junctions is further evident intheir 𝜂 values compared to modules with conventional contactlayers. As shown in Figure 4d, long-term thermal stability posesa significant challenge for Mg3(Bi,Sb)2 junctions. After aging,modules with Ni/Mg3(Bi,Sb)2 and 304SS/Mg3(Bi,Sb)2 junctionsexhibited significantly lower 𝜂 of 6.6% and 7.1%, respectively.To explore the potential for further performance enhance-ments, simulations were conducted to predict the module’s 𝜂,incorporating the performance of the TE materials, the 𝜌c ofboth n-type and p-type legs, and the properties of the TEiMs. De-tailed simulation results are presented in Figure S20 (Support-ing Information), showing that the simulated 𝜂max could reachas high as 11.8%. Due to the inherent challenges in accuratelymeasuring heat flow,[73] the experimentally measured Qc tendsto exceed the simulated values. To better illustrate the poten-tial of our developed module, the 𝜂 was recalculated using themeasured P (Pm) and the simulated heat flow (Qs). Under aΔT of 294 K, the recalculated 𝜂max reached an impressive 10.0%(Figure 4d). Compared to Bi2Te3-based and other Mg3(Bi,Sb)2-based modules (Figure 4e),[9,24,33,36,49,53,71,72] our module exhibitsa superior ɷmax, maintaining outstanding performance even af-ter 30 days of aging. Notably, the experimentally realized 𝜂max of8.6% surpasses most previously reported two-pair modules de-signed for low-grade heat harvesting applications (ΔT < 300 K)(Figure 4f).[9,24,33,36,49,53,71,72] Moreover, the outstanding thermalstability of Mg/Mg3(Bi,Sb)2 junctions highlights their promisefor long-term sustainable power generation. Looking ahead,the scalable fabrication of Mg3(Bi,Sb)2 materials remains a keychallenge for real-world applications. Mechanical alloying andmelting-sintering routes have shown promise for scaling upwhile preserving material performance.[74,75] Furthermore, theinterfacial design developed in this work is compatible with in-dustrial metallization techniques, such as chemical or electro-plating, commonly used in Bi2Te3-based TE modules.[76] Thesefactors highlight the feasibility of translating our findings towardpractical module-level manufacturing.3. ConclusionTo overcome the long-standing stability challenges at both thematerial and interface levels, we developed a holistically sta-ble Mg/Mg3(Bi,Sb)2 junction by leveraging activated Mg diffu-sion. Introducing a Mg interlayer between Mg3(Bi,Sb)2 and themetallization layer not only ensures ultralow 𝜌c but also preservesthe intrinsic TE properties of the material, achieving unprece-dented dual stability. TheMg/Mg3(Bi,Sb)2 junctionmaintains anultralow 𝜌c below 5 μΩ cm2 while sustaining high 𝜎 of the ma-terial even after aging for over 100 days at 573 K. The Mg layerserves as a robust diffusion barrier while simultaneously com-pensating for Mg loss at grain boundaries, thereby enhancingelectron transport within the TE material. This dual stability en-ables a remarkable 𝜂max of 8.6% and ɷmax of 0.45 W cm−2 undera ΔT of 294 K in a two-pair module, even after 30 days of agingof the n-type junctions. These findings highlight the exceptionalpotential ofMg/Mg3(Bi,Sb)2 junctions for sustainable power gen-eration in low-grade heat harvesting applications. Moreover, thedemonstrated dual stability in TE junctions provides valuable in-sights into achieving long-term reliability and enhanced perfor-mance in TE devices, paving the way for long-awaited practicaland sustainable TE power generation applications.4. Experimental SectionMaterials Synthesis: To prepare the Mg3.2In0.02Bi1.4Sb0.595Te0.005 (de-noted asMg3(Bi,Sb)2 in themain text), high-purity rawmaterials includingMg turnings (99.95%), Te shots (99.999%), Bi shots (99.999%), Sb shots(99.999%), and In powder (99.99%) were accurately weighed according tostoichiometry. The materials were loaded into a stainless-steel ball-millingjar inside an argon-filled glovebox. Ball milling was performed for 5 h ina single run using a SPEX 8000D high-energy mill. The resulting powderwas consolidated by spark plasma sintering (SPS, SPS-1080 System, SPSSYNTEX INC) under pressure of 60 MPa at 973 K for 10 min. The p-typeMgAgSb sample was prepared using a ball-milling process with the addi-tion of 0.625 wt.% C18H36O2. The obtained powder was consolidated bySPS at 573 K under 60 MPa for 5 min (SPS-322Lx, Dr. Sintering).Properties and Microstructure Characterization: The electrical proper-ties, including electrical conductivity (𝜎) and Seebeck coefficient (S), weremeasured using the ZEM-3 instrument in a helium atmosphere (AD-VANCE RIKO, ± 5% uncertainty). The thermal conductivity 𝜅 was calcu-lated using the equation 𝜅 = 𝜆Cpd,where the thermal diffusivity 𝜆wasmea-sured by a laser flash method (LFA 467, NETZSCH,± 3% uncertainty), thedensities d wasmeasured by an Archimedes method. The heat capacity Cpwas determined by the polynomial equation proposed by Agne et al.:[77]Cp = 3NR (1 + 1.3 × 10−4 T – 4 × 103 T−2) / Mw, where N is the num-ber of elements of the formula unit, Mw is the molecular weight of theformula unit. Microstructural and chemical compositions were character-ized using a field-emission scanning electronmicroscope (FESEM,HitachiSU8000) equipped with an energy dispersive spectrometer (EDS, XFlashFlatQUAD 5060F). Room-temperature contact resistance of the TE junc-tions was measured by a two-axis resistance distribution measurementinstrument (S1331, Mottainai energy).Module Fabrication and Efficiency Evaluation: To fabricate theNi/Mg/Mg3(Bi,Sb)2/Mg/Ni junction, Ni powder, Ni foil, Mg foil, andMg3(Bi,Sb)2 bulk were loaded into a graphene die with a sandwichstructure, followed by SPS at 773 K for 10 min. To minimize oxidation,freshly polished Mg foils were immediately transferred into an argon-filledglovebox for die assembly. Although the formation of a thin native MgOlayer is difficult to completely avoid, it is not expected to significantlycompromise the interfacial contact quality, as confirmed by low 𝜌c andmicrostructural analysis. The Ni/Mg3(Bi,Sb)2/Ni junction was fabricatedusing the same process, while the 304SS/Mg3(Bi,Sb)2/304SS junctionwas prepared by one-step SPS sintering under the same situation ofmaterial preparation. The obtained sandwich-structure legs were ground,polished and then cut into dice for module fabrication, with dimensions of3.8 mm × 3.8 mm × 6 mm for both n-type and p-type legs. The preparedlegs were placed inside quartz tubes and then evacuated for annealingin a muffle furnace. The Sb-based alloy was selected as the contact layerfor preparing p-type legs by one-step SPS sintering. Two-pair moduleswere fabricated based on these n-type and p-type legs. The electricaloutput power and generation performance of the fabricated moduleswere characterized using a commercial apparatus (Mini-PEM, ADVANCERIKO, Japan). The hot-side temperature (Th) of the modules was con-trolled by a heater, and the cold-side temperature (Tc) was maintained bythe flowing water. To characterize the performance of the TE junctionsafter aging, the junctions were re-soldered onto substrates. All moduleswere measured using the Mini-PEM system under identical operatingconditions to ensure consistency. The 3D finite-element simulationsof power-generation were performed with COMSOL Multiphysics®Adv. Mater. 2025, 37, 2508270 2508270 (8 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.dewww.advancedsciencenews.com www.advmat.desoftware. The model incorporated the Heat Transfer in Solids, ElectricCurrents, and Thermoelectric Effect interfaces, which were fully coupledvia the multiphysics node to capture the coupled thermal and electricaltransport within the devices. The simulated geometry replicated the actualtwo-pair TE module constructed in this work. The hot-side temperaturewas varied from 373 to 593 K, while the cold side was fixed at 295 K.To simplify the model, thermal contact resistances at the interfaces wasneglected. All material properties, including 𝜅, S, and 𝜎, were obtainedfrom experimental measurements.Supporting InformationSupporting Information is available from the Wiley Online Library or fromthe author.AcknowledgementsL.W. A.L. and X.W. contribute equally to this work. We acknowledge fi-nancial support from the JST Mirai Program Grant Number JPMJMI19A1.Center support from JSPS WPI Academy is also thanked.Conflict of InterestT.M. and L.W. have filed one Japanese patent application (2025-017528)on the work described here. The remaining authors declare no competinginterests.Data Availability StatementThe data that support the findings of this study are available from thecorresponding author upon reasonable request.Keywordselemental diffusion, interface engineering, Mg3(Bi, Sb)2, stability,thermoelectricsReceived: April 30, 2025Revised: June 5, 2025Published online: June 27, 2025[1] Y. Zhou, T. Ding, G. Xu, S. Yang, C.-W. Qiu, J. He, G. W. Ho, Nat. Rev.Phys. 2024, 6, 769.[2] J. D. Sachs, G. Schmidt-Traub, M. Mazzucato, D. Messner, N.Nakicenovic, J. Rockström, Nat. Sustain. 2019, 2, 805.[3] G. Schierning, Nat. Energy 2018, 3, 92.[4] G. J. Snyder, E. S. Toberer, Nat. Mater. 2008, 7, 105.[5] T. Mori, S. Priya,MRS Bull. 2018, 43, 176.[6] T. Hendricks, T. Caillat, T. Mori, Energies 2022, 15, 7307.[7] W. Liu, S. Bai, J. Materiomics 2019, 5, 321.[8] W. Liu, Q. Jie, H. S. Kim, Z. Ren, Acta Mater. 2015, 87, 357.[9] L. Xie, L. Yin, Y. Yu, G. Peng, S. Song, P. Ying, S. Cai, Y. Sun, W. Shi,H. Wu, N. Qu, F. Guo, W. Cai, H. Wu, Q. Zhang, K. Nielsch, Z. Ren,Z. Liu, J. Sui, Science 2023, 382, 921.[10] H. Tamaki, H. K. Sato, T. Kanno, Adv. Mater. 2016, 28, 10182.[11] J. Zhang, L. Song, S. H. Pedersen, H. Yin, L. T. Hung, B. B. Iversen,Nat. Commun. 2017, 8, 13901.[12] K. Imasato, S. D. Kang, G. J. Snyder, Energy Environ. Sci. 2019, 12,965.[13] J. Mao, H. Zhu, Z. Ding, Z. Liu, G. A. Gamage, G. Chen, Z. Ren, Sci-ence 2019, 365, 495.[14] X. Shi, T. Zhao, X. Zhang, C. Sun, Z. Chen, S. Lin, W. Li, H. Gu, Y. Pei,Adv. Mater. 2019, 31, 1903387.[15] R. Shu, Y. Zhou, Q. Wang, Z. Han, Y. Zhu, Y. Liu, Y. Chen, M. Gu, W.Xu, Y. Wang, W. Zhang, L. Huang, W. Liu, Adv. Funct. Mater. 2019, 29,1807235.[16] Z. Liu, N. Sato, W. Gao, K. Yubuta, N. Kawamoto, M. Mitome, K.Kurashima, Y. Owada, K. Nagase, C.-H. Lee, J. Yi, K. Tsuchiya, T. Mori,Joule 2021, 5, 1196.[17] N. Chen, H. Zhu, G. Li, Z. Fan, X. Zhang, J. Yang, T. Lu, Q. Liu, X. Wu,Y. Yao, Y. Shi, H. Zhao, Nat. Commun. 2023, 14, 4932.[18] J.-W. Li, Z. Han, J. Yu, H. L. Zhuang, H. Hu, B. Su, H. Li, Y. Jiang, L.Chen, W. Liu, Q. Zheng, J. F. Li, Nat. Commun. 2023, 14, 7428.[19] J. Lei, K. Zhao, J. Liao, S. Yang, Z. Zhang, T. R. Wei, P. Qiu, M. Zhu, L.Chen, X. Shi, Nat. Commun. 2024, 15, 6588.[20] P. Zhao, W. Xue, Y. Zhang, S. Zhi, X. Ma, J. Qiu, T. Zhang, S. Ye, H.Mu, J. Cheng, X. Wang, S. Hou, L. Zhao, G. Xie, F. Cao, X. Liu, J. Mao,Y. Fu, Y. Wang, Q. Zhang, Nature 2024, 631, 777.[21] A. Li, Y. Wang, Y. Li, X. Yang, P. Nan, K. Liu, B. Ge, C. Fu, T. Zhu, Nat.Commun. 2024, 15, 5108.[22] Y. Huang, J. Lei, H. Chen, Z. Zhou, H. Dong, S. Yang, H. Gao, T.-R.Wei, K. Zhao, X. Shi, Acta Mater. 2023, 249, 118847.[23] L. Xie, J. Yang, Z. Liu, N. Qu, X. Dong, J. Zhu, W. Shi, H. Wu, G. Peng,F. Guo, Y. Zhang, W. Cai, H. Wu, H. Zhu, H. Zhao, Z. Liu, J. Sui,Mater.Today 2023, 65, 5.[24] A. Li, L. Wang, J. Li, T. Mori, Energy Environ. Sci. 2024, 17, 8810.[25] Z. Liu, J. Mao, J. Sui, Z. Ren, Energy Environ. Sci. 2018, 11, 23.[26] Y. Wang, A. Li, Y. Hong, T. Deng, P. Deng, Y. Huang, K. Liu, J. Wang,C. Fu, T. Zhu,Nat. Mater. 2025, https://doi.org/10.1038/s41563-024-02112-7.[27] M. Liu, X. Zhang, S. Zhang, Y. Pei, Nat. Commun. 2024, 15, 6580.[28] Q. Yang, S. Yang, P. Qiu, L. Peng, T.-R. Wei, Z. Zhang, X. Shi, L. Chen,Science 2022, 377, 854[29] T. Cao, X. L. Shi, B. Hu, Q. Yang, W. Y. Lyu, S. Sun, L. C. Yin, Q. Y. Liu,W. Chen, X. Wang, S. Liu, M. Li, W. D. Liu, T. Tesfamichael, Q. Liu, J.MacLeod, Z. G. Chen, Nat. Commun. 2025, 16, 1555.[30] L. Yin, X. Li, X. Bao, J. Cheng, C. Chen, Z. Zhang, X. Liu, F. Cao, J.Mao, Q. Zhang, Nat. Commun. 2024, 15, 1468.[31] L. Miao, Q. Zhang, M. Yuan, R. Li, M. Wang, X. Tan, J. Wu, G. Q. Liu,J. Jiang, Adv. Mater. 2025, 2503580.[32] X. Wu, Z. Han, Y. Zhu, B. Deng, K. Zhu, C. Liu, F. Jiang, W. Liu, ActaMater. 2022, 226, 117616.[33] X. Wu, Y. Lin, Z. Han, H. Li, C. Liu, Y. Wang, P. Zhang, K. Zhu, F. Jiang,J. Huang, H. Fan, F. Cheng, B. Ge, W. Liu, Adv. Energy Mater. 2022,12, 2203039.[34] J. Chu, J. Huang, R. Liu, J. Liao, X. Xia, Q. Zhang, C. Wang, M. Gu, S.Bai, X. Shi, L. Chen, Nat. Commun. 2020, 11, 2723.[35] R. Liu, Y. Xing, J. Liao, X. Xia, C. Wang, C. Zhu, F. Xu, Z. G. Chen, L.Chen, J. Huang, S. Bai, Nat. Commun. 2022, 13, 7738.[36] Y. Fu, Q. Zhang, Z. Hu, M. Jiang, A. Huang, X. Ai, S. Wan, H. Reith,L. Wang, K. Nielsch, W. Jiang, Energy Environ. Sci. 2022, 15, 3265.[37] Y. Fu, X. Ai, Z. Hu, S. Zhao, X. Lu, J. Huang, A. Huang, L. Wang, Q.Zhang, W. Jiang, Nat. Commun. 2024, 15, 9355.[38] A. Li, L. Wang, J. Li, X. Wu, T. Mori, Nat. Commun. 2025, 16, 1502.[39] L. Borgsmiller, D. Zavanelli, G. J. Snyder, PRX Energy 2022, 1, 022001.[40] I. T. Witting, T. C. Chasapis, F. Ricci, M. Peters, N. A. Heinz, G.Hautier, G. J. Snyder, Adv. Electron. Mater. 2019, 5, 1800904.[41] J. Male, M. T. Agne, A. Goyal, S. Anand, I. T. Witting, V. Stevanovic,G. J. Snyder,Mater. Horiz. 2019, 6, 1444.[42] S. Ohno, K. Imasato, S. Anand, H. Tamaki, S. D. Kang, P. Gorai, H. K.Sato, E. S. Toberer, T. Kanno, G. J. Snyder, Joule 2018, 2, 141.[43] S. Bano, R. Chetty, J. Babu, T. Mori, Device 2024, 2, 100408.Adv. Mater. 2025, 37, 2508270 2508270 (9 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.dehttps://doi.org/10.1038/s41563-024-02112-7https://doi.org/10.1038/s41563-024-02112-7www.advancedsciencenews.com www.advmat.de[44] A. Li, C. Fu, X. Zhao, T. Zhu, Research 2020, 2020, 1934848.[45] H. Shang, Z. Liang, C. Xu, J. Mao, H. Gu, F. Ding, Z. Ren, Research2020, 2020, 1219461.[46] H. Cho, S. Y. Back, N. Sato, Z. Liu, W. Gao, L. Wang, H. D. Nguyen,N. Kawamoto, T. Mori, Adv. Funct. Mater. 2024, 34, 2407017.[47] K. Yang, X. Li, C. Sun, W. Song, W. Zhao, Q. Zhang, Adv. Funct. Mater.2024, 34, 2315886[48] H. Yang, B. Jia, L. Xie, D. Mao, J. Xia, J. Yang, M. Yuan, Q. Gan, X. Liu,M. Hu, J. Shuai, J. He, Joule 2024, 8, 1.[49] L. Wang, W. Zhang, S. Y. Back, N. Kawamoto, D. H. Nguyen, T. Mori,Nat. Commun. 2024, 15, 6800.[50] A. Li, P. Nan, Y. Wang, Z. Gao, S. Zhang, Z. Han, X. Zhao, B. Ge, C.Fu, T. Zhu, Acta Mater. 2022, 239, 118301.[51] J. Zhang, L. R. Jorgensen, L. Song, B. B. Iversen, ACS Appl. Mater.Interfaces 2022, 14, 31024.[52] H. Shang, Z. Liang, C. Xu, S. Song, D. Huang, H. Gu, J. Mao, Z. Ren,F. Ding, Acta Mater. 2020, 201, 572.[53] P. Ying, R. He, J. Mao, Q. Zhang, H. Reith, J. Sui, Z. Ren, K. Nielsch,G. Schierning, Nat. Commun. 2021, 12, 1121.[54] Z. Bu, X. Zhang, Y. Hu, Z. Chen, S. Lin, W. Li, C. Xiao, Y. Pei, Nat.Commun. 2022, 13, 237.[55] P. Ying, L.Wilkens, H. Reith, N. P. Rodriguez, X. Hong, Q. Lu, C. Hess,K. Nielsch, R. He, Energy Environ. Sci. 2022, 15, 2557.[56] C. Xu, Z. Liang, W. Ren, S. Song, F. Zhang, Z. Ren, Adv. Energy Mater.2022, 12, 2202392.[57] L. Yin, C. Chen, F. Zhang, X. Li, F. Bai, Z. Zhang, X. Wang, J. Mao, F.Cao, X. Chen, J. Sui, X. Liu, Q. Zhang, Acta Mater. 2020, 198, 25.[58] R. Chetty, J. Babu, T. Mori, ACS Appl. Energy Mater. 2024, 7, 12112.[59] Y. Wang, J. Chen, Y. Jiang, M. Ferhat, S. Ohno, Z. A. Munir, W. Fan, S.Chen, ACS Appl. Mater. Interfaces 2022, 14, 33419.[60] N. Qu, Y. Sun, Z. Liu, L. Xie, Y. Zhu, H. Wu, R. Chai, J. Cheng, F. Guo,Q. Zhang, R. Wang, B. Li, Z. Wei, W. Cai, J. Sui, Adv. Energy Mater.2023, 14, 2302818[61] J. Yang, G. Li, H. Zhu, N. Chen, T. Lu, J. Gao, L. Guo, J. Xiang, P. Sun,Y. Yao, R. Yang, H. Zhao, Joule 2022, 6, 193.[62] L.Wang, A. Li, J. Li, N. Kawamoto, D.H.Nguyen, T.Mori,Nano Energy2025, 137, 110797.[63] E. D. l. Barrera, Nat. Struct. Mol. Biol. 2005, 12, 280.[64] M. Liu, X. Zhang, J. Tang, Z. Chen, W. Li, Y. Pei, Sci. Bull. 2023, 68,2536.[65] T. V. Blank, Y. A. Gol’dberg, Semiconductors 2007, 41, 1263.[66] S. M. Sze, K. K. Ng, Physics of Semiconductor Devices, John Wiley &Sons, Hoboken, NJ, USA 2006.[67] H. Sepehri-Amin, K. Imasato, M. Wood, J. J. Kuo, G. J. Snyder, ACSAppl. Mater. Interfaces 2022, 14, 37958.[68] D. Li, F. Huang, B. Ren, S. Wang, W. Zhang, L. Zhu, Sci. Rep. 2024,14, 905.[69] Y. Geng, Z. Li, Z. Lin, Y. Liu, Q. Lai, X. Wu, L. Hu, F. Liu, Y. Yu, C.Zhang, Small 2023, 20, 2305670.[70] J. J. Kuo, Y. Yu, S. D. Kang, O. Cojocaru-Mirédin, M. Wuttig, G. J.Snyder, Adv. Mater. Interfaces 2019, 6, 1900429.[71] L. Yin, F. Yang, X. Bao, W. Xue, Z. Du, X. Wang, J. Cheng, H. Ji, J. Sui,X. Liu, Y. Wang, F. Cao, J. Mao, M. Li, Z. Ren, Q. Zhang, Nat. Energy2023, 8, 665.[72] M. Jiang, Y. Fu, Q. Zhang, Z. Hu, A. Huang, S. Wang, L. Wang, W.Jiang, Natl. Sci. Rev. 2023, 10, nwad095.[73] R. Chetty, J. Babu, T. Mori, Joule 2024, 8, 556.[74] Y. Liu, Y. Geng, Y. Dou, X. Wu, L. Hu, F. Liu, W. Ao, C. Zhang, Small2023, 19, 2303840.[75] Q. Zhang, M. Yuan, K. Pang, Y. Zhang, R. Wang, X. Tan, G. Wu, H.Hu, J. Wu, P. Sun, G. Q. Liu, J. Jiang, Adv. Mater. 2023, 35, 2300338.[76] Q. Zhang, Q. Pan, M. Wang, K. Pang, R. Li, L. Miao, X. Tan, H. Hu, J.Wu, P. Sun, G.-Q. Liu, J. Jiang, Acta Mater. 2025, 292, 121064.[77] M. T. Agne, K. Imasato, S. Anand, K. Lee, S. K. Bux, A. Zevalkink, A.J. E. Rettie, D. Y. Chung, M. G. Kanatzidis, G. J. Snyder,Mater. TodayPhys. 2018, 6, 83.Adv. Mater. 2025, 37, 2508270 2508270 (10 of 10) © 2025 The Author(s). Advanced Materials published by Wiley-VCH GmbH 15214095, 2025, 38, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/adma.202508270 by National Institute For, Wiley Online Library on [26/09/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advmat.de Active Diffusion Controlled Dual Stability in Thermoelectrics for Sustainable Heat Harvesting 1. Introduction 2. Results 2.1. Achievement of Dual Stability in Mg3(Bi,Sb)2 Junctions 2.2. Superior Stability of TE Materials in TE Junctions 2.3. Simultaneously High Stability of Contact in Mg/Mg3(Bi,Sb)2 Junctions 2.4. Blocking Mg Diffusion from Materials to the Contact Layer and Mg Compensation 2.5. Sustainable Power Generation in Modules using Mg/Mg3(Bi,Sb)2 Junctions 3. Conclusion 4. Experimental Section Supporting Information Acknowledgements Conflict of Interest Data Availability Statement Keywords