Field-like Torque in Spintronics

• Field-like torque is a component of spin torque that mimics an effective magnetic field along the spin polarization axis, influencing magnetization dynamics in nanostructures.
• It is characterized by the imaginary part of the spin-mixing conductance and is modulated by material, interface, and voltage effects in devices like MTJs and SOT systems.
• FLT critically tunes GHz oscillations, synchronization in spintronic nano-oscillators, and deterministic switching in memory devices, boosting overall performance.

Field-like torque (FLT) is a fundamental component of the current-induced spin torque acting on magnetization in layered magnetic nanostructures, such as spin valves, magnetic tunnel junctions (MTJs), and heavy-metal/ferromagnet (HM/FM) bilayers. FLT originates primarily from the transverse component of spin accumulation at magnetic interfaces and, unlike the Slonczewski or damping-like torque, mimics the action of an effective magnetic field oriented along the spin-polarization axis. Advances in both experimental quantification and theoretical modeling have established FLT as a key parameter for the dynamics, frequency tuning, synchronization, and deterministic switching in spintronic oscillators and memory devices.

1. Theoretical Foundations and Microscopic Origin

FLT appears in the generalized Landau-Lifshitz-Gilbert-Slonczewski (LLGS) equation as an additive torque term distinct from the damping-like (Slonczewski) torque. For a unit magnetization vector $\mathbf{m}$ acted on by a spin current with polarization $\mathbf{p}$, the spin-torque terms are conventionally decomposed: $$\frac{d\mathbf{m}}{dt} = -\gamma\,\mathbf{m}\times\mathbf{H}_\mathrm{eff} + \alpha\,\mathbf{m}\times\frac{d\mathbf{m}}{dt} + a_J\,\mathbf{m}\times(\mathbf{m}\times\mathbf{p}) + b_J\,\mathbf{m}\times\mathbf{p}$$ where $a_J$ is the amplitude of the damping-like (in-plane) torque and $b_J$ is the amplitude of the field-like torque (0810.3421, Fujimoto, 2022, Liu et al., 2021). FLT arises from the imaginary part of the spin-mixing conductance at HM/FM or MTJ interfaces, interfacial Rashba–Edelstein effects, as well as from Oersted fields in some configurations (Ou et al., 2016, Bose et al., 2017, Dutta et al., 2021).

Key features of the FLT:

• Mathematical form: $b_J\,\mathbf{m}\times\mathbf{p}$ (spin-transfer devices, perpendicular-to-plane transport), or $b_{\mathrm{SOT}}\,(\hat{p}\times\mathbf{m})$ for SOT systems.
• Physical effect: Acts as an effective field along $\mathbf{p}$, shifting the spontaneous axis of precession and modifying the overall energy landscape of the magnetization (Taniguchi et al., 2015, Taniguchi et al., 2014).
• Origin: Linked to transverse spin reflection and interfacial spin-orbit coupling (Fujimoto, 2022, Ou et al., 2016).

In spin Hall driven systems, FLT is proportional to the imaginary part of the spin-mixing conductance and can have either sign, depending critically on interfacial details and the symmetry of the spin system (Taniguchi, 2021, Liu et al., 2021).

2. Material Dependence, Amplitude, and Sign

FLT efficiency and sign vary sharply with materials choice, layer thickness, and interface engineering:

• Heavy-metal/ferromagnet systems: In Ir/CoFeB, the interfacial FLT torque conductivity can reach $$\sigma_\mathrm{FL}^\mathrm{int} \sim -5 \times 10^4$$ $\hbar/2e\,\Omega^{-1}$m$^{-1}$, about $36\%$ the value of the damping-like torque and with an opposite sign to the Oersted field; bulk contributions (spin Hall effect) add a smaller, typically opposite-sign component (Dutta et al., 2021).
• Cr/Ni bilayers: The sign of FLT reverses below a critical Cr thickness ($t_\mathrm{Cr}<6$ nm), attributed to the emergence of a strong interfacial Rashba–Edelstein effective field $\sim 35$ Oe per $10^{12}$ A/m$^2$ (Bose et al., 2017).
• Ru$_2$Sn$_3$/CoFeB: Unusually, thick Ru$_2$Sn$_3$ (10 nm) yields a large intrinsic FLT efficiency $\xi_\mathrm{FL}$ up to $-0.2$ at room temperature and $-0.29$ at $50$ K, greatly exceeding the damping-like torque and traceable to intrinsic Berry–curvature–driven spin Hall effect in the amorphous host (Peterson et al., 2020).
• Voltage bias in MTJs: In tunnel junctions, $b_J$ scales as $b_J \propto V^2$ (with $a_J \propto V$), with best-fit experimental coefficients in the range $\epsilon \sim 1.7$–$1.8$ V$^{-1}$ (0810.3421).

The magnitude and even the sign of FLT can be tuned through interface composition, thickness, spacer layers (to suppress interfacial Rashba contributions), oxidation, or material selection (Dutta et al., 2021, Bose et al., 2017, Ou et al., 2016).

3. Dynamical Consequences: Oscillations, Synchronization, and Energy Balance

FLT plays a pivotal role in stabilizing and tuning self-sustained oscillations in spin-torque and spin Hall oscillator devices:

• Stabilization of zero-field oscillations: For perpendicular-free-layer/in-plane-pinned oscillators, a moderate negative FLT is required for large-amplitude stable precession at zero applied field. Without FLT, no steady-state oscillation occurs; with FLT, the energy supplied by spin transfer torque can be exactly balanced by damping, leading to a stable limit cycle with frequency in the GHz range (Taniguchi et al., 2014, Taniguchi et al., 2015, Arun et al., 2021).
• Frequency and Q-factor tuning: The oscillation frequency and Q-factor (spectral purity) both increase with the magnitude of FLT (parameter $\beta$ or relative amplitude $n$). Exemplary frequency tunability of $2$–$80$ GHz with current or $\beta$ is reported in spin Hall oscillators with two in-plane easy axes (Arun et al., 2021). In coupled oscillator arrays, synchronization is strongly facilitated by FLT, which can lock large numbers of oscillators into phase and greatly boost collective microwave power output (Arun et al., 2023, Arun et al., 2019).
• Suppression of damping: At a critical amplitude $b_J = \alpha a_J$, FLT precisely cancels the Gilbert damping and triggers sustained auto-oscillations; further increase in FLT yields stable synchronized oscillations and minimizes multistability (Arun et al., 2019, Lakshmanan et al., 2021).
• Synchronization robustness: Arrays of serially coupled spin-torque nano-oscillators (STNOs) exhibit global synchronization above a threshold $\beta$ (often $\sim 0.1$–$0.3$); for very large arrays, field angle tuning is required in addition to FLT for full synchronization (Arun et al., 2023).

In field-theoretic analyses, FLT corresponds to a term $M_i \times m_j$ (cross product of magnetic moments in adjacent layers), and manifests in both quantum (field-theory/Kubo) and drift-diffusion models at a lower order in exchange coupling than damping-like torque, thus dominating under weak-coupling conditions (Fujimoto, 2022, Abert et al., 2016).

4. Functional Role in Magnetization Switching and Nonvolatile Memory

FLT modifies switching thresholds, speed, and determinism in spin torque- and spin–orbit-torque (SOT)-driven switching:

• Switching thresholds: A positive FLT lowers the critical switching current density in type-X SOT-MRAM geometries, accelerates energy dissipation, and broadens deterministic switching windows (Liu et al., 2021). In field-free switching with DMI, small FLT expands the operational current density window, while larger FLT can block switching due to modified domain wall dynamics (Wu et al., 2019).
• Oscillatory and unipolar switching: Large, oppositely signed FLT (relative to DLT) not only modulates domain wall chirality but can dynamically stabilize backward domain wall motion, enabling oscillatory and thus controllable unipolar deterministic switching using current pulse duration, without polarity inversion (Lee et al., 2017, Yoon et al., 2017).
• Suppression of precession ("ringing"): In perpendicular MRAM, large intrinsic FLT substantially reduces post-switching magnetization precession, yielding sharper, deterministic switching and lower write error rates (Peterson et al., 2020).
• Practical extraction: The amplitude of FLT can be extracted and quantified experimentally via protocols such as spin-torque FMR, second-harmonic Hall measurements, and easy-axis field sweeps, frequently validating macrospin theoretical predictions (Dutta et al., 2021, 0810.3421).

5. Temperature, Thickness, and Interface Effects

FLT exhibits strong dependence on thermal, geometric, and interfacial factors:

• Temperature: In systems where FLT is enhanced by spin-flip scattering at FM/oxide interfaces, its magnitude drops linearly with lowering temperature, vanishing below $\sim 70$ K as such scattering is frozen; the sign can revert to that of the HM/FM interface contribution at low $T$ (Ou et al., 2016).
• Thickness: FLT generally decays with increasing FM or HM thickness as the spin current generating it is depleted; in some materials the interfacial contribution persists for thicknesses below the spin diffusion length, whereas bulk contributions turn on slowly (Dutta et al., 2021).
• Interfacial engineering: Intercalated spacers (e.g., Cu in Cr/Cu/Ni) quench interfacial Rashba-driven FLT and restore the Oersted-dominated sign, providing direct control over torque sign and magnitude via heterostructure design (Bose et al., 2017).

6. Analytical and Experimental Parametrization

Quantitative characterization of FLT is essential for predictive modeling and device optimization:

Parameter/Quantity	Symbol	Characteristic Value/Dependence
Field-like torque amplitude	$b_J$ or $\tau_\mathrm{FL}$	$\propto V^2$ in MTJs; $\propto J_c$ in SOT
Relative strength (dimensionless)	$\beta$ or $n$	$0.01$–$0.3$ (experiment), up to $|n|\sim1$
Ratio FLT/DLT	$b_J/a_J$ or $n$	Up to $1$ in MTJs, $0.2$–$0.5$ in SOT devices
Effective field (macrospin)	$H_\mathrm{FL}$	$=b_J/\gamma$
Spin-torque efficiency	$\xi_\mathrm{FL}$	Varies; e.g., $-0.03$ in Ru$_2$Sn$_3$(4nm), $-0.20$ in Ru$_2$Sn$_3$(10nm) at 300K (Peterson et al., 2020)
Measurement protocols	—	ST-FMR, 2$\omega$ Hall, field sweeps

The voltage and current dependence, as well as the field and temperature scaling laws for FLT, have been robustly reproduced by both drift-diffusion and field-theoretical approaches (Fujimoto, 2022, Abert et al., 2016, Dutta et al., 2021).

7. Device Engineering and Practical Implications

FLT is now recognized as a critical parameter for the design and operation of spintronic circuits:

• Nano-oscillators: FLT enables GHz-band self-oscillation without external fields in STOs and SOT-based oscillators. Negative FLT can maximize power and Q-factor, while positive FLT optimizes frequency tunability (Arun et al., 12 Jan 2026, Arun et al., 2021).
• Memory devices: By tuning FLT, device engineers can lower writing current, optimize switching speed, and suppress stochastic delay in MRAM technologies (Liu et al., 2021, 0810.3421).
• SOT switching: Sizable positive FLT is advantageous in collinear (type-X) SOT-MRAM and domain-wall devices, enabling fast, deterministic switching at subnanosecond timescales and facilitating high-density integration (Liu et al., 2021).
• Synchronization: Arrays of STNOs with appropriate FLT can efficiently synchronize, boosting output power into the $\mu$W range for on-chip microwave applications and neuromorphic computing elements (Arun et al., 2023, Arun et al., 2019).
• Materials and interfaces: Engineering the sign and magnitude of FLT via barrier composition, interface quality, or layer thickness enables direct control over threshold, speed, and power in practical devices (Bose et al., 2017, Dutta et al., 2021, Peterson et al., 2020).

FLT, although sometimes subdominant to the damping-like torque, emerges as a non-negligible, often essential factor in the engineering of next-generation spintronic devices, enabling wide tunability, efficient synchronization, high Q-factors, and deterministic, low-power switching. Its quantitative understanding and control constitute a major area of contemporary spintronics research.