Permanent magnets are everywhere. They hold notes on refrigerators, drive motors in electric vehicles, and power the speakers in our headphones. But what if these ubiquitous magnets could also convert waste heat directly into electricity? Researchers have now developed a multifunctional composite magnet that does exactly that, achieving record-breaking performance in transverse thermoelectric generation while retaining full magnetic properties.
The innovation addresses a longstanding challenge in energy harvesting. Conventional thermoelectric devices, which convert temperature differences into electrical power, rely on the Seebeck effect and require complex architectures with substrates, electrodes, and thermal barriers that limit efficiency. Transverse thermoelectric effects offer a simpler geometry by generating electric current perpendicular to the applied heat flow, eliminating many structural complications. Yet previous attempts to combine magnetic functionality with high thermoelectric performance have fallen short, hampered by poor material properties or inefficient designs.
The new multifunctional composite magnet, or MCM, changes that equation. By alternately stacking layers of SmCo5 permanent magnet and Bi0.2Sb1.8Te3 thermoelectric material at a carefully optimized angle, the research team achieved a transverse thermoelectric figure of merit of 0.20 at room temperature. That performance is more than two orders of magnitude better than earlier magnetic thermoelectric materials and approaches levels seen in commercial devices based entirely on the Seebeck effect.
Engineering Heat Flow at the Nanoscale
The key to the MCM lies in a phenomenon called the off-diagonal Seebeck effect, which arises in artificially tilted multilayers when materials with contrasting electrical and thermal properties are stacked at an angle. When a temperature gradient is applied, heat does not flow uniformly through the composite. Instead, it preferentially travels along paths of high thermal conductivity, creating a nonuniform heat current. Because the layers are tilted, this nonuniform flow generates a transverse electric field through the ordinary Seebeck effect operating within each layer.
The effect is amplified when the two materials have opposite types of charge carriers. SmCo5, a well known permanent magnet prized for its strength and thermal stability, behaves like an n-type metal with high electrical and thermal conductivity and a negative Seebeck coefficient. Bi0.2Sb1.8Te3, a p-type thermoelectric semiconductor, has much lower conductivity but a positive Seebeck coefficient. When these materials are combined in alternating layers tilted at 25 degrees, the transverse electric currents generated in each layer add up rather than canceling out.
Previous studies predicted that such configurations could yield high transverse thermoelectric performance, but experimental results consistently fell short of theoretical expectations. The problem was interfacial resistance. At the boundaries between metal and semiconductor layers, electrical and thermal resistance can balloon, choking off current flow and dissipating heat before it can be converted into useful power. Most earlier work either ignored these interfaces or failed to control them.
The research team tackled this problem head-on by using spark plasma sintering to bond SmCo5 disks and Bi0.2Sb1.8Te3 powder under high pressure and moderate temperature. This process created thin interfacial diffusion layers composed of cobalt telluride compounds, which act as metallic adhesive bonds with electrical properties similar to SmCo5 itself. Detailed scanning electron microscopy and energy dispersive X-ray spectroscopy revealed that these interfacial layers are only about 10 micrometers thick and contain no voids or elemental migrations that would degrade performance.
To quantify the interfacial contributions, the researchers measured electrical resistance across the SmCo5 and Bi0.2Sb1.8Te3 multilayer by scanning a probe along the stacking direction. The resistance profile showed smooth, step-like increases corresponding to the bulk resistivities of each material, with almost no discontinuity at the interfaces. The interfacial electrical resistivity was calculated at just 0.4 micro-ohm square centimeters, comparable to the best contact resistivities achieved in conventional thermoelectric devices. This value is negligibly small, accounting for only about 1.2 percent of the total volumetric resistance.
Thermal resistance at the interfaces was characterized using lock-in thermography, a technique that detects temperature modulations induced by alternating current. When current flows through a junction with finite thermal resistance, heat absorption and release create localized temperature changes that can be visualized with an infrared camera. The measurements showed no detectable jumps in temperature or phase at the SmCo5 and Bi0.2Sb1.8Te3 boundaries, indicating that interfacial thermal resistance is below one millionth of a square meter kelvin per watt. That contribution is less than 0.4 percent of the total thermal resistance, confirming that the interfaces are essentially transparent to both charge and heat flow.
Record Power Generation From a Transverse Device
With interfacial losses minimized, the MCM delivered on its theoretical promise. Direct measurements of the off-diagonal Seebeck coefficient and electrical resistivity confirmed a transverse thermoelectric figure of merit of 0.20 at room temperature, rising to 0.32 at 420 kelvin. This performance surpasses all previous transverse thermoelectric materials by a wide margin and enters the realm of practical application.
To demonstrate real-world potential, the team constructed a lateral thermopile module by stacking 14 MCM elements with alternating tilt angles and connecting them electrically in series. This zigzag architecture, first proposed decades ago but rarely implemented successfully, allows the transverse voltages from each element to add up while maintaining a compact form factor. The fill factor, the fraction of the heat transfer area occupied by active thermoelectric material, exceeded 90 percent thanks to the transverse geometry and ultrathin insulating layers.
When subjected to a temperature difference of 152 kelvin, the module generated an open circuit voltage of 219 millivolts and delivered a maximum output power of 204 milliwatts. Normalized by heat transfer area and the square of the temperature gradient, the power density reached 0.17 milliwatts per square centimeter per squared kelvin per millimeter, the highest value ever reported for a transverse thermoelectric device. Remarkably, this performance is also competitive with commercial longitudinal thermoelectric modules based on bismuth telluride, which have been optimized over decades of development.
The module retained its magnetic properties throughout testing. The remanent magnetization of the SmCo5 layers measured 0.86 tesla, with a coercivity of 0.87 tesla, values typical of high quality permanent magnets. The magnetization remained stable above 600 kelvin, reflecting the excellent thermal stability of SmCo5. In a demonstration of multifunctionality, the magnetized module easily held several metal paper clips, confirming that the MCM operates simultaneously as both a permanent magnet and a thermoelectric generator.
Estimated conversion efficiency for the module ranged from 1.6 to 2.4 percent at a temperature difference of 152 kelvin. While modest compared to mechanical heat engines, this efficiency is record-breaking for transverse thermoelectrics and achieved without the complex substrate and electrode assemblies that burden conventional devices. The absence of hot side junctions also eliminates a common failure mode in longitudinal modules, where thermal stress and diffusion degrade contacts over time.
Magnetic Attraction Improves Thermal Contact
The multifunctionality of the MCM extends beyond simultaneous magnetism and thermoelectric conversion. The magnetic attractive force provides a practical advantage for energy harvesting applications. When placed on a ferromagnetic heat source, such as a steel plate or engine block, the MCM module self-adheres, eliminating air gaps and reducing thermal contact resistance without the need for clamps or adhesives.
To test this effect, the researchers prepared three versions of an eight element thermopile module: one demagnetized without a built-in heat sink, one magnetized without a heat sink, and one magnetized with a built-in heat sink formed by varying the height of neighboring elements to increase surface area for air cooling. All three were placed on a heated stainless steel plate and cooled by ambient air flow, simulating a realistic energy harvesting scenario.
The magnetized module without a heat sink generated twice the open circuit voltage of the demagnetized version at the same plate temperature, demonstrating that magnetic adhesion significantly enhances thermal coupling. Adding the built-in heat sink increased voltage by another 10 percent by improving heat dissipation on the cold side. Maximum output power scaled accordingly, reaching substantially higher values for the magnetized module with optimized thermal design.
This result highlights a broader principle: integrating magnetic functionality into thermoelectric materials can improve system-level performance even when the intrinsic material properties remain unchanged. The magnetic force ensures intimate contact with heat sources, reduces assembly complexity, and enables passive cooling architectures that would be impractical with conventional modules.
Why Transverse Thermoelectrics Matter
Transverse thermoelectric effects have long intrigued researchers because of their geometric advantages. Unlike longitudinal devices, which require alternating p-type and n-type legs connected by metallic bridges, transverse devices can operate as single, monolithic elements. This simplifies fabrication, eliminates electrode-substrate thermal interfaces, and allows larger temperature gradients to be applied directly to the active material.
Despite these benefits, transverse thermoelectrics have struggled to compete with established Seebeck effect devices. The ordinary Nernst effect, the oldest known transverse mechanism, requires strong external magnetic fields and delivers tiny voltages. The anomalous Nernst effect in magnetic materials operates without external fields but suffers from extremely low figures of merit, typically below 0.001. Composite structures based on the off-diagonal Seebeck effect promised better performance but were hindered by poor material choices and uncontrolled interfaces.
The MCM breakthrough resolves these obstacles by combining a high performance thermoelectric semiconductor with a practical permanent magnet and engineering near-perfect interfaces between them. The result is a device that matches or exceeds the power density of commercial modules while offering unique installation and thermal management advantages.
The choice of SmCo5 as the magnetic component was deliberate. Unlike the neodymium iron boron magnets used in an earlier attempt, SmCo5 has a negative Seebeck coefficient opposite in sign to that of p-type bismuth antimony telluride. This alignment of charge carrier types is essential for the off-diagonal Seebeck effect to produce additive rather than canceling voltages. SmCo5 also has excellent high temperature stability, maintaining magnetization well above 600 kelvin, and its metallic conductivity ensures low electrical resistance in the composite.
The interfacial engineering proved equally critical. During spark plasma sintering, cobalt and tellurium atoms diffuse across the boundary to form a thin intermetallic layer. Rather than acting as a barrier, this layer serves as a conductive adhesive that mechanically bonds the SmCo5 and Bi0.2Sb1.8Te3 while preserving low electrical and thermal resistance. The formation of this favorable interfacial phase appears to be a fortunate consequence of the materials' chemical compatibility under the chosen sintering conditions.
Future Directions and Broader Impact
The demonstrated performance of the MCM opens multiple avenues for further development. One straightforward path is to explore other permanent magnet materials with higher Seebeck coefficients or lower resistivities than commercial SmCo5. The researchers suggest that hybridizing the off-diagonal Seebeck effect with other transverse mechanisms, such as the anomalous Nernst effect or the Seebeck driven anomalous Hall effect, could further boost performance. Materials with large anomalous transport coefficients might enable figures of merit exceeding 0.5, approaching the best longitudinal thermoelectrics.
From a practical standpoint, the MCM concept is immediately applicable wherever permanent magnets are already deployed. Electric motors, magnetic bearings, and holding fixtures all generate or dissipate heat during operation. Integrating thermoelectric functionality into these components could enable self-powered sensors, wireless monitoring systems, or localized cooling without adding bulk or complexity. The magnetic attractive force simplifies retrofitting existing equipment, and the robust transverse geometry resists mechanical stress and thermal cycling.
The lateral thermopile architecture demonstrated here also suggests new module designs. By varying element height to create built-in heat sinks, designers can optimize thermal management for specific applications without relying on external hardware. The high fill factor and absence of ceramic substrates reduce material costs and weight, important considerations for portable or embedded systems.
Energy conversion efficiency remains a target for improvement. The 1.6 to 2.4 percent efficiency achieved at a 152 kelvin temperature difference is respectable for a first-generation transverse device but lags behind the best longitudinal modules, which can exceed 10 percent efficiency under optimized conditions. However, the junctionless structure of transverse devices offers potential advantages in durability and high temperature operation that may offset lower efficiency in demanding environments.
The MCM also illustrates a broader trend in materials science: the convergence of multiple functionalities into single materials or structures. Rather than optimizing for one property at the expense of others, the MCM achieves high performance in both magnetism and thermoelectricity by careful design of composition, microstructure, and interfaces. This multifunctional approach could extend to other material systems, combining mechanical, optical, or electronic properties in ways that enable entirely new device concepts.
A New Class of Functional Material
The development of the multifunctional composite magnet represents a significant milestone in thermoelectric research. By achieving a transverse thermoelectric figure of merit of 0.20 and demonstrating practical power generation with record power density, the work validates decades of theoretical predictions and overcomes longstanding experimental challenges. The near-elimination of interfacial resistances, the retention of strong magnetic properties, and the successful integration into a thermopile module all point toward real world viability.
Perhaps most importantly, the MCM concept challenges the assumption that permanent magnets must be passive components. By adding direct heat-to-electricity conversion, these materials gain a second function that could enable autonomous sensors, distributed energy harvesting, and new approaches to thermal management. As energy efficiency becomes increasingly critical across industries, multifunctional materials like the MCM offer a path toward doing more with less.
The journey from laboratory curiosity to commercial product is never short, but the MCM has cleared several critical hurdles. The materials are commercially available, the fabrication process is scalable, and the performance metrics are competitive with established technologies. Whether in automobiles, factories, or consumer electronics, the ability to harvest waste heat directly from magnetized components could become a routine feature of next generation systems. For now, the MCM stands as proof that permanent magnets can do more than attract and repel. They can also power the future.
Credit & Disclaimer: This article is a popular science summary written to make peer-reviewed research accessible to a broad audience. All scientific facts, findings, and conclusions presented here are drawn directly and accurately from the original research paper. Readers are strongly encouraged to consult the full research article for complete data, methodologies, and scientific detail. The article can be accessed through https://doi.org/10.1039/d4ee04845h
