Magnon-Drag Thermopiles: Mechanisms & Devices

• Magnon-drag thermopiles are devices that utilize temperature gradients to drive magnon flows in magnetic materials, transferring momentum to charge carriers to produce a measurable voltage.
• Engineered architectures like pairwise arrays and hybrid junctions isolate the magnon-drag signal from conventional electron and phonon contributions through magnetic control.
• Theoretical frameworks, including Boltzmann transport and spin-motive force models, quantitatively describe magnon-electron interactions to guide material selection and device optimization.

A magnon-drag thermopile is a thermoelectric device in which thermal gradients drive magnon flows in magnetic materials, and the resulting magnon dynamics impart momentum or energy to conduction electrons (or holes), generating a voltage via magnon-drag mechanisms. These devices exploit the collective magnetic excitations—magnons—in ferromagnets, antiferromagnets, or ferrimagnets, and are engineered to maximize the magnon-drag Seebeck coefficient, as distinct from conventional electron-diffusive or phonon-drag terms. The magnon-drag thermopile concept is directly linked to the microscopic theory of magnon-electron and magnon-magnon scattering, and is realized through specially designed arrays of magnetic elements wired electrically in series and thermally in parallel, enabling direct, quantitative measurement and utilization of magnon-drag thermopower—often isolated from other thermoelectric effects by symmetry, magnetic configuration, or interfacial engineering (Costache et al., 2012, Prakash et al., 2018).

1. Fundamental Principles of Magnon-Drag Thermopiles

The essential physics underlying magnon-drag thermopiles is the transfer of momentum between magnons (bosonic spin-wave excitations) and mobile charge carriers in a magnetic system subject to a temperature gradient. When a thermal gradient is applied, magnons are driven from the hot to the cold end, and their interactions with electrons (or holes) can "drag" the charge carriers, creating an additional thermoelectric voltage over and above diffusive or phonon-drag contributions. The corresponding Seebeck coefficient can be written in the general form:

$$S_\text{total}(T) = S_\text{el}(T) + S_\text{ph}(T) + S_\text{mag}(T)$$

where $S_\text{el}$ is electron-diffusion, $S_\text{ph}$ is phonon-drag, and $S_\text{mag}$ is the magnon-drag term (Costache et al., 2012, Prakash et al., 2018).

The magnitude and sign of $S_\text{mag}$ depend on the nature of the magnon spectrum, the electron-magnon (or magnon-carrier) coupling strength, and the relaxation rates of the relevant scattering mechanisms. Typical temperature dependencies—arising from Boltzmann transport, Kubo or Landau-Lifshitz-Gilbert (LLG) theory—are $S_\text{mag} \propto T^{3/2}$ in 3D ferromagnets with a quadratic magnon dispersion (Flebus et al., 2016, Miura et al., 2012, Watzman et al., 2016, Yamaguchi et al., 2018).

2. Device Architectures and Experimental Isolation of Magnon-Drag

There are several canonical device geometries for magnon-drag thermopiles, chosen to maximize sensitivity to magnon-drag while suppressing non-magnetic background contributions:

• Pairwise Thermopile Arrays: Arrays of ferromagnetic nanowires are arranged in pairs between a hot and cold reservoir, each pair having distinct coercivities, allowing magnetic configuration control (parallel vs antiparallel). Electrically in series and thermally in parallel, the configuration enables clean isolation of the magnon-drag Seebeck coefficient via magnetic field switching—ordinary electron and phonon-drag contributions cancel by design (Costache et al., 2012).
• Hybrid Differential Junctions: Heterostructures such as Pt|YIG and Pt|GGG are patterned into U-shaped or serpentine arrays, with the voltage difference measured between arms on different substrates under the same thermal bias. This direct differential measurement yields the difference in interfacial Seebeck coefficients and allows precise determination of magnon impedance effects (Prakash et al., 2018).
• Multilayer and Series-Stacked Architectures: Alternating magnetic thin films and spacers (e.g., in FePt/MgO or Heusler alloy/insulator stacks) are electrically connected in series, thermally in parallel, greatly amplifying the magnon-drag-induced voltage through geometric scaling (Yan et al., 2013, Matsuura et al., 2021).
• Interfaces with Magnetic and Nonmagnetic Conductors: In magnetic insulator/metal bilayers (e.g., YIG|Pt), the interfacial magnon Hall effect and drag by topological surface magnons can be detected and exploited via the induced voltage in the adjacent metal (De et al., 2022).

These architectures allow the systematic separation of magnon-drag from electronic or phononic contributions, enable control via external magnetic fields, and facilitate scaling to multijunction thermopiles for enhanced output.

3. Theoretical Descriptions and Quantitative Models

Magnon-drag thermopower in thermopiles can be captured by several frameworks:

1. Boltzmann Transport and Hydrodynamic Models: Here, electrons and magnons are treated as coupled fluids exchanging momentum. In the hydrodynamic regime and the limit where magnon-electron scattering dominates, the magnon-drag Seebeck coefficient is

$$S_\text{mag}(T) \approx \frac{2}{3} \frac{C_m(T)}{n_e e}$$

where $C_m$ is the magnon heat capacity and $n_e$ is the electron density (Watzman et al., 2016).

2. Spin-Motive-Force and Berry-Phase Mechanisms: In the presence of dynamic magnetization textures, the spin-motive force contributes an additional voltage, with a term proportional to material-specific parameters such as the Gilbert damping $\alpha$ and the dissipative spin-transfer parameter $\beta$:

$$S_\text{md} = \left( \frac{\beta}{3\alpha} - 1 \right) J\left(\frac{\lambda}{\xi}\right) \frac{k_B P}{\pi^2 e} \left(\frac{T}{T_c}\right)^{3/2}$$

where $P$ is the spin polarization and $J(\lambda/\xi)$ a dimensionless factor related to the magnon spectrum (Flebus et al., 2016, Lucassen et al., 2011).

1. Microscopic Kubo and Green’s Function Approaches: Detailed evaluations of $L_{12}$ and $L_{11}$ transport coefficients (thermoelectric and electrical conductivities) in generalized linear-response frameworks yield the Seebeck coefficient as $S = L_{12}/(T L_{11})$; magnon-drag emerges via electron-magnon correlation functions (Yamaguchi et al., 2018, Matsuura et al., 2021).
2. Phonon and Magnon-Drag in Multi-Subband or Noncollinear Systems: In ferrimagnets and canted AFMs, quartic and cubic magnon-magnon interactions generate notable inter- or intra-band drag. The resulting enhancement (or suppression) in spin and heat conductivities, and associated Seebeck effects, is strongly dependent on the detailed magnon band structure, exchange coupling, and field-dependent canting angles (Arakawa, 2022, Arakawa, 2022).

These models accommodate both the regime where the magnon-drag is positive (electron drag toward hot end) and negative (Umklapp-dominated drag, with electrons dragged toward cold end) (Yan et al., 2013).

4. Materials Selection and Optimization Criteria

Effective magnon-drag thermopiles require materials and interfaces where magnon-drag is both strong and controllable:

• Ferromagnetic Metals and Alloys: Permalloy, Fe, Co, Ni, and ordered phases such as L1₀-FePt, Fe₂V₀.₈W₀.₂Al (Heusler alloys), are prime candidates due to sizable magnon populations, suitable exchange stiffness, and tunable band structures. Large spin polarization and strong electron-magnon coupling are critical for maximizing $S_\text{mag}$ (Costache et al., 2012, Miura et al., 2012, Watzman et al., 2016, Matsuura et al., 2021, Yan et al., 2013).
• Magnetic Insulators: YIG (yttrium iron garnet) enables studies of pure magnon transport and magnon Hall phenomena, often in multilayer sandwich structures with heavy metals such as Pt or W for ISHE voltage readout and interface engineering (Prakash et al., 2018, Lyapilin et al., 2017, De et al., 2022).
• Antiferromagnets and Ferrimagnets: Systems such as REFeAsO, canted antiferromagnets, and ferrimagnetic insulators show strong electron-magnon or interband magnon-magnon drag effects, especially near the magnon entropy peak below the ordering transition (Caglieris et al., 2014, Arakawa, 2022, Arakawa, 2022).
• Semiconducting Magnets: Doped MnTe and related compounds allow realization of magnon-drag thermopower in p- and n-type configurations with controllable sign and magnitude, essential for p–n magnon-drag thermocouples (Polash et al., 2020).

Relevant material properties are optimized via doping (to tune carrier concentration and band alignments), control of impurity and defect concentrations (to maximize magnon lifetimes and reduce parasitic scattering), and engineering of spin-orbit interactions (to control $\beta$ and band spin-mixing) (Yan et al., 2013, Matsuura et al., 2021).

5. Practical Device Scaling, Performance, and Engineering Guidelines

Magnon-drag thermopiles can be scaled in arrayed or stacked configurations, where performance metrics such as open-circuit voltage, power density, and Seebeck coefficient scale linearly or quadratically with the number of junctions or stages under ideal series connection:

• Voltage Scaling: Total Seebeck coefficient $S_\text{total} = N S_\text{single}$ for $N$-element series arrays. For example, with $S_\text{single} \sim 8\,\mu$V/K for a Pt|GGG vs. Pt|YIG MDT at 8 K, a 100-element array gives $S_\text{total} \sim 0.8$ mV/K (Prakash et al., 2018, Costache et al., 2012).
• Power Output: The maximum power is $$P_\text{max} = (S_\text{total} \Delta T)^2/(4 R_\text{total})$$, with $R_\text{total}$ the internal resistance and $A$ the area. For thin-film systems, power densities of order $\mu$W/cm$^2$ are realizable at modest temperature gradients (Prakash et al., 2018, Matsuura et al., 2021).
• Device Geometry: Optimal devices use U-shaped, serpentine, or parallel-series wire arrays for compactness and thermal uniformity, and are fabricated using standard lithography techniques. Magnetic configuration and thickness (e.g., YIG layer tuning from 40–250 nm) allow tuning of magnon impedance and drag amplitude (Costache et al., 2012, Prakash et al., 2018).
• Thermal Management and Contacts: Use of high-quality interfaces (e.g., low-resistance Pt, Au, or Ag), dielectric spacers, and careful heat-sink design is mandatory for minimizing parasitic losses and maintaining device performance under finite thermal bias (Matsuura et al., 2021).
• Control Parameters: Device performance is tunable by external magnetic fields (to "freeze out" or enhance magnon populations), choice of magnetic layer thickness, and modification of spin-mixing conductance at interfaces (Prakash et al., 2018, Lyapilin et al., 2017, De et al., 2022).

A sample summary table of experimental Seebeck coefficients, voltages per junction, and scaling is as follows:

Structure (Material)	$S_\text{mag}$ per element (µV/K)	Array size $N$	$S_\text{total}$ (mV/K)
Pt	GGG vs. Pt	YIG MDT	8
Permalloy wires (20 nm)	$\sim20$ (at peak)	20	0.4
FePt (10 nm) (Umklapp)	$-3$ to $-30$ (var. $T$)	100	$-0.3$
Heusler-alloy leg	$-500$ (max)	100	$-50$

6. Representative Results and Major Experimental Findings

• Ferromagnetic Metal Arrays: Clear isolation of magnon-drag with precisely tunable signal via control of wire magnetization; observation of a $T^{3/2}$ scaling of the drag coefficient; confirmation of relaxation-time asymmetry (drag maximal when $\tau_{m,e} \ll \tau_{m,x}$), consistent with fundamental predictions (Costache et al., 2012, Watzman et al., 2016).
• Hybrid Metal/Insulator MDTs: Demonstration that the introduction of a ferromagnetic insulator (YIG) at the metal interface impedes phonon drag, and that YIG film thickness and applied field modulate the magnon-drag signal with suppression or recovery of drag peaks correlating with the magnon scattering volume (Prakash et al., 2018).
• Heusler Alloy Magnon-Drag: Achievement of record Seebeck coefficients (up to $-500$ µV/K) near 300 K in W-doped Fe$_2$VAl films, with magnon-drag dominating over diffusive contributions; realization of projected output voltages of several volts and sub-watt power output in large-scale thermopile stacks (Matsuura et al., 2021).
• Antiferromagnetic and Bipolar-Drag Systems: Systematic exploration of magnon-drag in p–n semiconductor systems (e.g., Cr-doped MnTe) reveals sign changes and relaxation-time independence in the bipolar regime, enabling engineered p–n magnon-drag junctions with tunable voltage output (Polash et al., 2020).
• Topological and Noncollinear Magnon Systems: Observation and modeling of drag effects mediated by topological surface magnons or cubic magnon-magnon interactions can yield strong, field-enhanced magnon-drag peaks an order of magnitude larger than simple noninteracting predictions (De et al., 2022, Arakawa, 2022).

7. Limitations, Challenges, and Outlook

Several factors limit the practical deployment and optimization of magnon-drag thermopiles:

• Temperature Range: Peak magnon-drag typically occurs at cryogenic to intermediate temperatures, determined by the magnon gap and underlying ordering temperature.
• Relaxation Mechanisms: At higher temperatures, magnon-magnon and magnon-phonon scattering suppress the drag; at very low T, magnon population vanishes.
• Scalability and Fabrication: High-density integration of thermocouple legs, controlled interface quality, and uniformity of thin-film microstructure are technologically demanding.
• Thermal Conductivity: High lattice or magnonic thermal conductivity can reduce the thermoelectric figure of merit $ZT$ below that of state-of-the-art phonon-drag or quantum-well devices; minimizing parasitic heat flow is crucial (Matsuura et al., 2021, Caglieris et al., 2014).
• Contact and Substrate Losses: Interfacial contact resistance and substrate heat leakage must be minimized for efficient energy harvesting.
• Material Optimization: Open questions persist around enhancing drag via band engineering (Umklapp, SOC-mixing), maximizing magnon lifetimes, and optimizing spin-mixing conductance at metal-magnet interfaces (Yan et al., 2013, Prakash et al., 2018).

Magnon-drag thermopiles—by leveraging advanced materials science, interfacial engineering, and precise magnetic control—now provide a powerful platform for exploring spin-caloritronic transport, energy harvesting at low temperatures, and fundamental magnon-electron and magnon-magnon interaction physics. The continued convergence of experimental and theoretical developments enables rational, predictive design of magnon-drag-based thermoelectric devices with new functionalities and record Seebeck coefficients (Costache et al., 2012, Prakash et al., 2018, Matsuura et al., 2021, Watzman et al., 2016, Yan et al., 2013).