The rapidly growing demand for thermal management in advanced fields such as integrated circuits, optical communications, and low-temperature medicine has accelerated the development of efficient, quiet, and compact thermoelectric cooling devices^[1-3]. However, conventional Bi₂Te₃-based systems are constrained by the scarcity of tellurium (Te) and limited performance at elevated temperatures due to their narrow band gap and associated bipolar effects, which lead to rapid efficiency degradation above about 400 K^[4,5]. In contrast, Mg₃(Sb, Bi)₂ exhibits excellent thermoelectric performance over a broad temperature range (300-773 K), along with advantages such as low cost, light weight, and good mechanical robustness^[6,7]. These attributes make them strong candidates to replace n-type Bi₂Te₃ alloys in next-generation devices^[8]. However, their practical application is severely hindered by a critical challenge: the high chemical reactivity of Mg makes these materials vulnerable to moisture-induced corrosion, leading to rapid performance degradation^[9,10]. Existing protection strategies, such as using organic, HfO₂, Al₂O₃, and Mg-Mn alloy coatings, provide only partial mitigation and remain susceptible to failure due to cracking or delamination^[11,12]. More importantly, they fail to address corrosion during material storage and processing. Therefore, developing a fundamental solution that ensures intrinsic humidity stability while preserving high thermoelectric performance throughout the entire material lifecycle remains an urgent priority.

In a study published in Nature Materials, Sui and co-workers introduced an innovative sacrificial anodic protection method into the design of Mg₃(Sb, Bi)₂ materials and devices^[13]. Through multi-objective optimization, they screened and constructed in situ uniformly distributed multi-scale Mg₁₇Al₁₂ anode second phases within the Mg₃(Sb, Bi)₂ matrix [Figure 1]. Owing to its lower equilibrium potential, Mg₁₇Al₁₂ forms micro-galvanic cells that preferentially corrode, thereby providing thermodynamic cathodic protection to the matrix. Meanwhile, corrosion-induced Mg/Al hydroxides and oxides generate a self-healing hybrid passivation layer, which further isolates the material from moisture. The self-healing capability depends on the Pilling-Bedworth (PB) ratio^[14]. A PB ratio slightly above unity facilitates self-healing by generating a moderate compressive stress that ensures newly formed oxides seamlessly seal surface voids, preventing further moisture infiltration. This intrinsic material design reduces the average corrosion rate of Mg₃(Sb, Bi)₂ by 92% (to 95 μm·year^-1) in air and by 86% (to 0.36 μm·h^-1) in water, effectively solving corrosion issues during material storage and processing that cannot be addressed by conventional coating strategies. Importantly, since the Mg₁₇Al₁₂ phase is generated in situ by introducing Al, the content of Mg in the matrix will also be concomitantly influenced. Thus, the content of Mg vacancies, as well as the carrier concentration and thermoelectric properties, might change with the addition of Al^[15]. For practical applications, we must balance the stability and energy-conversion efficiency. In the Nature Materials study^[13], the thermoelectric performance is even slightly enhanced by incorporating 20 at% Al. Moreover, the biased carrier concentration in the matrix due to the in situ formation of the second phase can be further optimized via tuning the content of other dopants, such as Te and lanthanide elements^[16]. Therefore, the second-phase engineering, in conjunction with carrier concentration optimization, can enable an ideal balance between stability and functionality. This strategy is also generalizable, as demonstrated in other moisture-sensitive systems, including Mg₂(Sn, Ge) and CaMg₂Bi₂, through incorporating Al to construct anodic secondary phases. Yet, the appropriate dopants can vary from Al, depending on the matrix material, the thermodynamically favorable second phase, and the difference in equilibrium potential between them. The Nature Materials article offers an excellent paradigm and principles for screening suitable element candidates.

In addition to material optimization, Sui and colleagues tackled another hidden corrosion source that limits device-level reliability, namely accelerated material corrosion induced by the interface barrier layer. The commonly used Fe interface layer possesses a much higher equilibrium potential than the Mg₃(Sb, Bi)₂ matrix, accelerating corrosion of the matrix during operation^[17]. To overcome this, the team replaced Fe with Mg₁₇Al₁₂ as the interface layer. This design maintains low contact resistivity, matched thermal expansion, and chemical stability, which are essential requirements for thermoelectric interface materials^[18,19], while simultaneously eliminating interfacial galvanic corrosion and providing cathodic protection. As a result, no performance degradation of the thermoelectric device was observed after 28 days under harsh hot-humid conditions (350 K, 70% RH). The study also reminds us that the galvanic stability, which had been overlooked, should also be included in the design principles of thermoelectric interface materials in future studies.

This work shifts the design paradigm for thermoelectric materials and devices from conventional coating strategies to sacrificial anode protection. It highlights the power of an integrated, multi-scale approach combining electrochemistry, materials science, and device engineering to enhance both stability and functionality. Beyond enabling moisture-resistant and high-performance Mg-based thermoelectric devices, this study establishes a broadly applicable design strategy for corrosion-resistant functional materials.