Field-like torque

For decades, controlling magnetism meant applying an external magnetic field, a force that causes a magnet's orientation to precess, much like a spinning top wobbling under gravity. However, the advent of spintronics has introduced a revolutionary alternative: using spin-polarized electrical currents to exert sophisticated control over magnetic states. This has opened up new avenues for technology, but also raised fundamental questions about the nature of the interaction between electrical currents and magnetization. The core challenge is to understand and categorize the complex torques that arise from this interaction.

This article addresses this challenge by deconstructing the concept of current-induced spin torques, focusing specifically on the ​​field-like torque​​. It unpacks how these torques, despite their complex quantum origins, can be elegantly classified into two fundamental types dictated by the profound principles of symmetry and energy conservation.

In the following chapters, you will gain a deep understanding of this crucial concept. The "Principles and Mechanisms" chapter will define the field-like torque, contrasting it with its dissipative counterpart, the damping-like torque. We will explore how symmetry dictates their form, how they relate to the flow of time and energy, and what unifies them at a microscopic level. Subsequently, the "Applications and Interdisciplinary Connections" chapter will demonstrate the practical importance of this torque, from techniques for its precise measurement to its role in engineering futuristic memory devices and its surprising connections to fields like superconductivity and mechanics.

Think about a simple toy top spinning on a table. If you give it a gentle nudge, it doesn't just fall over. Instead, it begins a slow, graceful wobble—a precession. A magnet is much like that top. Its "spin" is its magnetization, a vector we can call $\mathbf{m}$. When you place it in a magnetic field, $\mathbf{H}$, the magnet doesn't simply align with the field. It precesses around it. The "nudge" from the field is a torque, a twisting force, described by the elegant cross product $\boldsymbol{\tau} \propto \mathbf{m} \times \mathbf{H}$. For decades, this was the primary way we knew to control magnetism: apply a magnetic field. But the world of spintronics has unveiled a new, more subtle, and often more powerful way to command magnets: using electrical currents.

It turns out that a current of electrons, if their intrinsic spins are aligned, can exert torques on a magnet. This spin-polarized current acts as a gust of "spin wind" that can make the magnetization precess, switch, or even oscillate at gigahertz frequencies. The beauty of it all is that despite the complex quantum mechanics at play, these new torques can be understood by classifying them into just two fundamental types, whose forms are not arbitrary but are profoundly dictated by the symmetries of the physical world.

Let’s imagine a spin-polarized current, carrying spins aligned along a direction $\boldsymbol{\sigma}$, interacting with a ferromagnet whose magnetization is $\mathbf{m}$. The resulting torque isn't just one simple thing; it's a combination of two distinct actions.

The first, and perhaps most intuitive, is the ​​field-like torque​​. It has the mathematical form: $\boldsymbol{\tau}_{\mathrm{FL}} \propto \mathbf{m} \times \boldsymbol{\sigma}$ Look closely at this expression. It has the exact same structure as the torque from a magnetic field, $\mathbf{m} \times \mathbf{H}$. The spin polarization $\boldsymbol{\sigma}$ is playing the role of an effective magnetic field. This is why we call it "field-like." It's a conservative, or ​​reactive​​, torque. Just like a magnetic field, it causes the magnetization $\mathbf{m}$ to precess around the direction of $\boldsymbol{\sigma}$ without causing it to lose energy.

The second player on this stage is the ​​damping-like torque​​. Its form is a bit more complex, involving a double cross product: $\boldsymbol{\tau}_{\mathrm{DL}} \propto \mathbf{m} \times (\mathbf{m} \times \boldsymbol{\sigma})$ This vector lies in the plane defined by $\mathbf{m}$ and $\boldsymbol{\sigma}$ and tries to push $\mathbf{m}$ either toward or away from the direction of $\boldsymbol{\sigma}$. It acts much like air resistance on our spinning top. A spinning top eventually slows down and spirals toward an upright position due to friction, or damping. This torque can either enhance or counteract the natural damping in the magnet. It is a ​​dissipative​​ torque—it changes the magnet's energy.

These two torques are the fundamental building blocks for describing how currents control magnets. The complete modern equation of motion for magnetization, the generalized Landau-Lifshitz-Gilbert (LLG) equation, incorporates both. It describes the precession due to effective fields, the natural Gilbert damping that makes the precession spiral inward, and these two new spin torques that allow us to drive the system: $\frac{d\mathbf{m}}{dt} = - \gamma \mathbf{m}\times \mathbf{H}_{\mathrm{eff}} + \alpha \mathbf{m} \times \frac{d\mathbf{m}}{dt} + \boldsymbol{\tau}_{\mathrm{FL}} + \boldsymbol{\tau}_{\mathrm{DL}}$ Here, $\gamma$ is the gyromagnetic ratio, and $\alpha$ is the Gilbert damping parameter. This equation is the workhorse of modern magnetism, and the field-like and damping-like torques are its spintronic engine.

You might wonder, why these two specific forms? Are there other possibilities? The answer, beautifully, is no. The very existence and form of these torques are dictated by symmetry, one of the most powerful and profound principles in physics.

Consider a typical device structure: a thin layer of a heavy metal (like platinum or tungsten) topped with a thin ferromagnet (like cobalt or permalloy). This bilayer has a clear up-down asymmetry—the top is different from the bottom. We can represent this broken inversion symmetry with a vector pointing normal to the interface, let’s call it $\hat{\mathbf{z}}$. However, within the plane of the layers, the material is often isotropic; you can rotate it without changing its properties. This is a system with $C_{\infty v}$ symmetry.

Now, let's pass an electrical current through the heavy metal layer, say along the $\hat{\mathbf{x}}$ direction. How can this situation produce a torque on the ferromagnet? According to Neumann's Principle, any physical response must respect the symmetry of the system. The current $\mathbf{J}$ (a polar vector along $\hat{\mathbf{x}}$) and the interface normal $\hat{\mathbf{z}}$ (a polar vector) are the only defining directions. The only unique direction you can produce from these two that is different from them is via the cross product: $\hat{\mathbf{z}} \times \mathbf{J}$. Since $\mathbf{J}$ is along $\hat{\mathbf{x}}$, this stimulus must be along the $\hat{\mathbf{y}}$ direction. So, any effective field or spin polarization generated must point along $\hat{\mathbf{y}}$.

Given this stimulus vector, let's call it $\boldsymbol{\sigma} \propto \hat{\mathbf{y}}$, what are the possible torques on the magnetization $\mathbf{m}$? A torque must be perpendicular to $\mathbf{m}$. The two simplest, linearly independent vectors you can construct from $\mathbf{m}$ and $\boldsymbol{\sigma}$ that are perpendicular to $\mathbf{m}$ are precisely $\mathbf{m} \times \boldsymbol{\sigma}$ and $\mathbf{m} \times (\mathbf{m} \times \boldsymbol{\sigma})$. And there you have it! Symmetry alone demands that any torque linear in the current must be a combination of a field-like and a damping-like component. They aren't just a convenient choice; they are the only choice the universe allows in this geometry.

The names "field-like" and "damping-like" hint at a deeper distinction related to energy. Let’s make this connection rigorous by considering the fundamental symmetry of time-reversal.

Imagine a frictionless pendulum swinging. If you film it and play the movie backward, the motion looks perfectly natural. The force of gravity, which depends on position, is a ​​reactive​​ or conservative force. Now, imagine a pendulum swinging through molasses. It quickly slows down. The drag force, which depends on velocity, is ​​dissipative​​. If you play that movie backward, you see something impossible: a pendulum in molasses spontaneously starting to swing higher and higher, with the molasses kicking it along. The dissipative drag force doesn't obey time-reversal symmetry.

Torques in magnetism follow the same classification. A reactive torque should be "time-reversal even"—it should look the same if time flows forward or backward. A dissipative torque should be "time-reversal odd"—it must flip its sign to make physical sense in a time-reversed world. Let's check our spin torques. Under time reversal ($T$), magnetization and spin polarization, being angular momenta, flip sign: $\mathbf{m} \to -\mathbf{m}$ and $\boldsymbol{\sigma} \to -\boldsymbol{\sigma}$.

For the field-like torque: $T(\boldsymbol{\tau}_{\mathrm{FL}}) \propto T(\mathbf{m} \times \boldsymbol{\sigma}) = (-\mathbf{m}) \times (-\boldsymbol{\sigma}) = \mathbf{m} \times \boldsymbol{\sigma} = +\boldsymbol{\tau}_{\mathrm{FL}}$ It is time-reversal even. It is reactive.

For the damping-like torque: $T(\boldsymbol{\tau}_{\mathrm{DL}}) \propto T(\mathbf{m} \times (\mathbf{m} \times \boldsymbol{\sigma})) = (-\mathbf{m}) \times ((-\mathbf{m}) \times (-\boldsymbol{\sigma})) = (-\mathbf{m}) \times (\mathbf{m} \times \boldsymbol{\sigma}) = -\boldsymbol{\tau}_{\mathrm{DL}}$ It is time-reversal odd. It is dissipative.

This is not just a formal classification. It has direct consequences for energy. Since the field-like torque acts just like a magnetic field, it can be derived from a potential energy term. It can do work on the magnet, but this work is stored as potential energy and can be recovered. It cannot, over a full cycle of precession, cause a net loss or gain of energy. It is non-dissipative. The damping-like torque, however, cannot be derived from a potential. It continuously pumps energy into (or extracts it from) the magnetic system, leading to dissipation. So, if you want to use a current to switch a magnet's direction, you need a dissipative torque—a damping-like torque—to overcome the system's own damping and push it over an energy barrier. The field-like torque can help shape the trajectory, but the damping-like torque does the heavy lifting of changing the system's energy.

Another key symmetry is inversion of the magnetization, $\mathbf{m} \to -\mathbf{m}$. The field-like torque is odd ($\boldsymbol{\tau}_{\mathrm{FL}} \to -\boldsymbol{\tau}_{\mathrm{FL}}$), while the damping-like torque is even ($\boldsymbol{\tau}_{\mathrm{DL}} \to +\boldsymbol{\tau}_{\mathrm{DL}}$) under this operation,. This difference allows experimentalists to separate and measure them.

So where do these torques come from? The answer lies in the quantum mechanical dance of electrons at interfaces. Let's look at one beautiful mechanism: the Rashba-Edelstein effect. At an asymmetric interface, the electric potential gradient creates a special kind of spin-orbit coupling. When a current flows, this coupling forces passing electrons to align their spins. For a current in the $\hat{\mathbf{x}}$ direction, a net spin accumulation $\mathbf{s}_E$ appears, pointing in the $\hat{\mathbf{y}}$ direction.

This stationary cloud of polarized spins acts like an effective magnetic field on the ferromagnet's magnetization $\mathbf{m}$, producing a purely field-like torque, $\boldsymbol{\tau}_{\mathrm{FL}} \propto \mathbf{m} \times \mathbf{s}_E$. But the story doesn't end there. The spins in this cloud now feel the enormous internal exchange field of the ferromagnet itself, and they begin to precess around $\mathbf{m}$. As they precess, they generate a new component of spin polarization, one that rotates with $\mathbf{m}$. It is this precessed component of spin that, when absorbed by the magnet, gives rise to the damping-like torque. In this picture, the FL torque is the primary effect, and the DL torque emerges as its precessional "echo."

This hints at a deep connection, which can be made more general. The interaction at the interface can be described by a complex ​​spin-mixing conductance​​, $G^{\uparrow\downarrow}(\omega) = G_r(\omega) + iG_i(\omega)$.

• The real part, $G_r$, represents dissipative processes—real spin-flips where angular momentum is irreversibly transferred. It governs the damping-like torque.
• The imaginary part, $G_i$, represents reactive processes—coherent precession and quantum phase shifts where no spin-flip occurs. It governs the field-like torque.

Here comes the final, elegant twist. In any physical system, an effect cannot precede its cause. This principle of ​​causality​​ imposes a rigid mathematical constraint on any response function like $G^{\uparrow\downarrow}(\omega)$. It implies that its real and imaginary parts are not independent. They are linked by the ​​Kramers-Kronig relations​​. If you know the full frequency dependence of the dissipative part ($G_r$), you can uniquely calculate the reactive part ($G_i$), and vice-versa. This means the field-like torque is not a separate, independent phenomenon. It is the causal, reactive counterpart to the dissipative damping-like torque. They are two faces of a single, unified spin-transfer process.

The profound unity of physics often reveals itself in unexpected connections. We saw that the Rashba effect, caused by a broken inversion symmetry at an interface, can generate both field-like and damping-like torques when a current flows. This is a dynamic effect. But what does this same symmetry breaking do to the magnet at rest, in equilibrium?

It gives rise to another fascinating phenomenon: the ​​Dzyaloshinskii-Moriya interaction (DMI)​​. The DMI is an antisymmetric exchange interaction that adds a twist to magnetism. It makes neighboring spins prefer to be slightly canted with respect to each other, rather than perfectly parallel. It is the origin of fascinating magnetic objects like skyrmions, which are tiny, stable magnetic whirls.

The crucial insight is that both the field-like spin-orbit torque and the interfacial DMI stem from the very same underlying physics: the spin-orbit coupling at the structurally asymmetric interface. They are both proportional to the strength of the Rashba effect. This shared origin leads to a powerful prediction: any change to the interface that modifies the DMI should modify the field-like torque in a corresponding way. For example, if you invert the interface by swapping the layers (e.g., from Pt/Co to Co/Pt), you flip the sign of the structural asymmetry. This is known to reverse the sign of the DMI constant. Because of their common origin, the field-like torque coefficient must also reverse its sign. This stunning prediction, born from symmetry arguments, has been confirmed by experiments, providing a powerful testament to the deep, unifying principles that govern the world of magnetism and spin. The force that twists a magnet at rest is intimately related to the torque that drives it into motion.

Now that we have grappled with the principles and mechanisms of the field-like torque, you might be tempted to ask, "So what?" Is this just another esoteric term in a physicist's equation, a subtle correction to be calculated and then perhaps forgotten? The answer, you will be delighted to find, is a resounding no! This torque is not a footnote; it is a headline. It is a key that unlocks new technologies, a lens that reveals hidden features of the quantum world, and a bridge between seemingly disparate realms of physics.

In this chapter, we will embark on a journey to see where this concept takes us—from the workbench of the experimentalist to the drawing boards of future computers, and into the very heart of condensed matter physics, where magnetism intersects with mechanics, superconductivity, and the profound art of symmetry.

Before we can apply a physical concept, we must first be convinced of its existence. We cannot "see" a torque directly, but we can be clever and observe its unmistakable effects on the world. The field-like torque, despite its quantum origins, leaves very tangible fingerprints on magnetic systems.

Imagine trying to flip a magnetic bit in a memory device. You apply an external magnetic field, pushing the magnetization from "up" to "down," then reverse the field to push it back. The field values at which the flips occur define a "hysteresis loop"—a signature of the magnet's memory. Now, let’s perform the same experiment while passing a spin-polarized electric current through the magnet. Suddenly, the entire loop shifts. It is no longer centered at zero magnetic field. It behaves as if a ghostly, invisible magnetic field is constantly present, helping the external field flip the switch in one direction and hindering it in the other. That "ghost" is the field-like torque, acting precisely as its name suggests: like a magnetic field. By simply measuring the shift of the hysteresis loop's center, experimentalists can quantify the strength of this otherwise invisible force.

This static measurement is elegant, but we can also make the system dance to see the torque in action. In a powerful technique called Spin-Torque Ferromagnetic Resonance (ST-FMR), we drive the magnet with a microwave-frequency current. This alternating current generates oscillating torques that try to make the magnetization precess, much like a spinning top. Think of pushing a child on a swing. A well-timed push at the end of the arc adds energy and increases the amplitude; this is analogous to the damping-like torque. The field-like torque is something stranger. It’s like momentarily shortening and lengthening the swing's ropes in time with the motion. It doesn't add energy in the same way, but it changes the phase of the swing relative to your pushes. In the ST-FMR experiment, these two types of "pushes" create a resonance signal with a unique and complex shape as we vary an external magnetic field. By carefully analyzing this shape—separating its symmetric (energy-giving) and antisymmetric (phase-shifting) components—physicists can precisely measure both torques at once, often using sophisticated fitting algorithms to extract the values from real data.

Of course, the real world is messy. Like a detective trying to listen to a whisper in a crowded room, a physicist must filter out a cacophony of spurious signals. The very current that generates our desired torques also heats the material, creating thermal voltages that can masquerade as the signal we are looking for. Here, physicists deploy their most powerful tool: symmetry. They know that the true torque signal has a specific "fingerprint." For instance, what happens if we could flip the magnet's north and south poles? It turns out the field-like and damping-like torques respond differently to this flip—one is "odd" under this transformation, while the other is "even". And what if we reverse the main magnetic field holding the magnet in place? The true torque signals respond in one way, while the thermal impostors respond in another. By cleverly combining measurements taken with opposite current and field directions, we can make the unwanted background noise cancel itself out, revealing the pure, unadulterated signal of the torques we seek. It is a beautiful demonstration of how abstract principles become a profoundly practical tool for discovery.

Once we can reliably measure and control these torques, we can put them to work building the next generation of electronics.

One of the most exciting ideas is "racetrack memory," where data is stored not as stationary bits, but as a series of magnetic boundaries—"domain walls"—that can be shuttled at high speeds along a nanowire. To move these walls, we need an engine. Spin-orbit torques are the perfect engine. The damping-like torque provides a powerful push, but in many advanced materials, the field-like torque plays a crucial and subtle role. It can help overcome internal forces that try to lock the domain wall in place, acting in concert with other quantum effects like the Dzyaloshinskii-Moriya interaction to achieve unprecedented speeds and efficiencies. The race to the future of data storage is being run on these tiny magnetic racetracks, powered by the very torques we have been discussing.

The next frontier is perhaps even more exotic: antiferromagnets. In these materials, neighboring atomic magnets point in opposite directions, perfectly canceling each other out. This makes them incredibly fast, robust against external fields, and they produce no stray fields to interfere with their neighbors—ideal for ultra-dense devices. But this stability is also a curse: how do you control something that is so perfectly balanced? A uniform magnetic field won't do. The answer, once again, lies in the field-like torque. In special antiferromagnets whose crystal lattice has a beautiful "inversion-partner" symmetry, applying a uniform electric current produces a staggered field-like torque. Imagine two gears spinning in opposite directions. This effect acts like an effective magnetic field that points "up" on one gear and "down" on the other. This action, perfectly tailored to the antiferromagnetic order, efficiently twists the entire system in unison and switches its state. This remarkable effect, born from the crystal's own symmetry, provides the key to unlocking the technological promise of these materials.

Perhaps the greatest beauty of a fundamental concept in physics is its power to connect phenomena that, at first glance, seem to have nothing to do with one another. The field-like torque is a prime example, weaving a thread through magnetism, superconductivity, and even mechanics.

Consider the bizarre realm of superconductivity, where electrons pair up and flow as a quantum fluid with zero resistance. What happens when this frictionless current passes through a thin magnetic barrier in a device known as a Josephson junction? It, too, becomes spin-polarized and can exert a field-like torque on the magnet, with a strength that can be tuned by the quantum phase difference across the junction. This is an extraordinary unification, a meeting point for the collective spin ordering of magnetism and the macroscopic quantum coherence of superconductivity.

The connections become even more surprising. What does a magnet have to do with sound? A sound wave traveling through a crystal is, at its heart, a rhythmic mechanical vibration of atoms. This mechanical strain can couple to the spins and induce an effective field-like torque, literally making the magnetic order dance to the tune of a sound wave. This link between magnetism and mechanics—a field known as magneto-acoustics—also reveals a profound truth about the nature of the field-like torque itself. Because it acts like a field (producing a torque $\boldsymbol{\tau} \propto \mathbf{L} \times \mathbf{H}_{\text{eff}}$), its action is always perpendicular to the magnetization's motion. As a result, it does no work over a closed cycle. Much like the force of gravity on a planet in a perfectly circular orbit, which continuously changes the planet's direction but not its speed, the field-like torque steers the magnetization without "pushing" it to dissipate energy. It is a truly conservative, or "reactive," interaction.

From a simple curiosity in an equation to a practical tool for measurement and engineering, and finally, to a unifying principle echoing across physics, the story of the field-like torque is a testament to the interconnected beauty of the natural world.