Kittel Formula

Ferromagnetic resonance (FMR) is a fundamental phenomenon in magnetism, describing the collective resonant absorption of microwave energy by a magnet's aligned spins. While seemingly esoteric, this effect is a cornerstone for both fundamental research and cutting-edge technology. However, a key challenge lies in understanding and predicting the precise resonance frequency, which mysteriously shifts depending on the magnet's shape and composition. This article addresses this gap by providing a comprehensive guide to the Kittel formula, the elegant theoretical framework that brings clarity to this complex behavior. In the following chapters, we will first explore the "Principles and Mechanisms," dissecting the physics of spin precession and the crucial roles of demagnetizing fields and anisotropy that culminate in the Kittel formula. Subsequently, under "Applications and Interdisciplinary Connections," we will see how this powerful tool is applied in practice to characterize materials and drive innovations in fields ranging from spintronics to quantum computing.

Now that we have a sense of what ferromagnetic resonance is, let’s peel back the layers and look at the beautiful physics humming away underneath. How does it work? Why is the resonance frequency what it is? You might think that trying to describe the collective dance of billions upon billions of atomic magnets would be a hopelessly complicated affair. And yet, as we’ll see, a few elegant principles bring a stunning clarity to the whole picture.

Imagine a simple spinning top. If you try to tip it over, it doesn’t just fall. Instead, it begins a slow, graceful wobble, or ​​precession​​, around the vertical direction of gravity. The Earth’s gravitational pull exerts a torque that tries to pull it down, but because the top is spinning, this torque instead nudges its axis of rotation sideways.

A magnetic moment—the tiny quantum arrow of magnetism at the heart of an electron—behaves in much the same way. When you place it in a magnetic field, the field exerts a torque on it, trying to align it perfectly. But due to its intrinsic quantum "spin," which is a form of angular momentum, the magnetic moment doesn't just snap into alignment. Instead, it precesses around the direction of the magnetic field, just like the spinning top. This fundamental dance is described by a beautifully compact law of nature, the ​​Landau-Lifshitz equation​​:

This equation is the rulebook for our magnetic dance. It says that the rate of change of the magnetization, $\mathbf{M}$, is dictated by the torque generated by its interaction with an ​​effective magnetic field​​, $\mathbf{H}_{\text{eff}}$. The constant $\gamma$ is the gyromagnetic ratio, a fundamental property of the electron that sets the tempo of the precession. The key to everything that follows lies in figuring out what this "effective" field, the field the magnetization actually feels, consists of.

Let's start our journey with the most perfect shape imaginable: a sphere. Suppose we have a ferromagnetic sphere, meaning all the little atomic moments are aligned, creating a net saturation magnetization, $M_s$. We place it in an external magnetic field, $H_0$. What is the resonance frequency?

You might brace for a complicated answer. After all, the magnetization within the sphere itself generates its own magnetic field, a so-called ​​demagnetizing field​​, which adds to the external field. This self-generated field arises because the "north poles" on one side of the sphere and "south poles" on the other create a field that passes back through the material, opposing the magnetization that created it.

So the effective field, $\mathbf{H}_{\text{eff}}$, is the sum of the external field we apply, $\mathbf{H}_0$, and this internal demagnetizing field, $\mathbf{H}_d$. For a sphere, a wonderful simplification occurs due to its perfect symmetry. The demagnetizing field inside a uniformly magnetized sphere is uniform and always points exactly opposite to the magnetization, with a strength proportional to it: $\mathbf{H}_d = -N_s \mathbf{M}$, where $N_s$ is the demagnetizing factor for a sphere.

Now, let’s see what happens. The total effective field is $\mathbf{H}_{\text{eff}} = \mathbf{H}_0 - N_s \mathbf{M}$. When we excite the system, the magnetization vector $\mathbf{M}$ starts to precess in a small cone around the external field $\mathbf{H}_0$. As $\mathbf{M}$ wobbles, its own demagnetizing field $\mathbf{H}_d$ wobbles right along with it, always pointing perfectly opposite. This internal field adds a constant "drag" that tries to pull the magnetization back to the axis, but it does not provide the kind of sideways nudge that would change the frequency of the precession itself. When you solve the Landau-Lifshitz equation for this case, the terms involving the magnetization and its own field miraculously cancel out of the dynamics of the frequency. The result is astonishingly simple:

The resonance frequency of the entire sphere is exactly the same as it would be for a single, isolated spin! The complex internal interactions, in this case of perfect symmetry, conspire to have no effect on the resonance frequency. It’s a beautiful example of how symmetry can simplify a seemingly complex problem.

But the real world is rarely made of perfect spheres. Much of our modern technology, from computer memory to sensors, is built on ​​thin films​​. And for a thin film, the story changes dramatically.

Imagine a film lying flat in the x-y plane. Its demagnetizing field is now highly anisotropic—it cares a great deal about direction. If you try to magnetize the film in its plane (say, along the x-axis), the north and south poles are very far apart, and the resulting demagnetizing field is nearly zero. But if you try to magnetize it perpendicular to the plane (along the z-axis), you are squashing the north and south poles close together on the large flat surfaces. This creates a powerful demagnetizing field inside the film, $H_d = -M_s$, that fiercely opposes the perpendicular magnetization. This shape-induced preference for in-plane magnetization is a form of ​​shape anisotropy​​.

How does this affect the resonance? Let's consider two experiments, just as a researcher might perform in a lab:

1. ​​Field Applied In-Plane ($\mathbf{H}_0$ along x-axis):​​ The magnetization is happily aligned in its preferred plane. We give it a small kick. To precess, it must develop small components in the y and z directions. Wobbling in the y-direction (still in-plane) is easy. But wobbling in the z-direction (out-of-plane) is very hard, because the strong demagnetizing field immediately pulls it back. The magnetization is tracing an elliptical path, and the restoring forces are different in different directions. This asymmetry changes the dynamics. The resonance frequency is no longer simple; it becomes a geometric mean of the effective fields in the two directions of motion:

   $$\omega_{\parallel} = \gamma \sqrt{H_0 (H_0 + M_s)}$$

   The term $(H_0 + M_s)$ reflects the strong restoring force for out-of-plane motion.

2. ​​Field Applied Perpendicular ($\mathbf{H}_0$ along z-axis):​​ First, we need a very strong external field ($H_0 > M_s$) just to overcome the demagnetizing field and force the magnetization to point out of the plane. Once it's aligned, the effective static field it feels is the external field minus the opposing demagnetizing field: $H_{\text{eff,z}} = H_0 - M_s$. The precession now happens in the x-y plane, where the demagnetizing fields are negligible. The dynamics become simple again, but based on this reduced effective field:

   $$\omega_{\perp} = \gamma (H_0 - M_s)$$

Look at these two results! By simply changing the orientation of the same film in the same external field, we get completely different resonance frequencies. The geometry of the sample is not a minor detail; it is a central character in the story, fundamentally altering the internal magnetic environment.

We've seen the sphere and the thin film. You might be wondering: is there a master formula that can handle any simple shape? For a general ellipsoid, the answer is a resounding yes. Charles Kittel provided this masterstroke, now known as the ​​Kittel formula​​.

For an ellipsoid with its principal axes along x, y, and z, and the external field $H_0$ applied along z, the resonance frequency is given by:

Here, $N_x, N_y,$ and $N_z$ are the ​​demagnetizing factors​​, dimensionless numbers that depend only on the aspect ratios of the ellipsoid. They perfectly encode the geometry. This single, powerful equation contains all our previous results as special cases:

• For a ​​sphere​​, $N_x = N_y = N_z$, so the terms in parentheses vanish, and we recover $\omega_0 = \gamma H_0$.
• For a ​​thin film​​ with the field in-plane (e.g., along x), we relabel the axes so the field is along z. The film is thin in the y-direction, so $N_x \approx 0, N_z \approx 0$, but $N_y \approx 1$. Plugging this into the formula (after careful axis relabeling) gives the in-plane result.
• For a ​​thin film​​ with the field perpendicular (along z), we have $N_x \approx 0, N_y \approx 0, N_z \approx 1$. The formula becomes $\omega_0 = \gamma \sqrt{(H_0 - M_s)(H_0 - M_s)} = \gamma (H_0 - M_s)$.

This is the kind of unity physicists dream of. One elegant expression, rooted in the simple idea of precession, can predict the dynamic behavior of magnets in a vast range of geometries.

Shape is not the only thing that dictates a magnet's behavior. The atoms in a crystal are arranged in a specific lattice, and this underlying structure can create preferred directions for magnetization. This is called ​​magnetocrystalline anisotropy​​. It's an intrinsic property of the material, like its color or density.

How do we account for this? We can use the powerful concept of the effective field again! We imagine that the crystal lattice creates its own internal magnetic field, an ​​anisotropy field​​ $H_K$, that helps to hold the magnetization along a preferred "easy axis."

If we have a thin film with an in-plane easy axis and we apply the external field $H_0$ along that same axis, the total effective field that holds the magnetization in place is strengthened to $H_0 + H_K$. When this is incorporated into the Kittel formula, the resonance frequency becomes:

This formula shows that the resonance frequency is sensitive not just to the field we apply, but also to these subtle, internal properties of the material. This is not just a theoretical curiosity; it's the basis of a powerful experimental technique. By measuring the resonance frequency at different applied fields, experimentalists can work backwards and determine the values of both the saturation magnetization $M_s$ and the anisotropy field $H_K$ with great precision, allowing them to characterize new materials for future technologies.

So far, we have treated the saturation magnetization $M_s$ as a fixed, constant property. But at any temperature above absolute zero, the magnetic order is not perfect. The atoms in the material are jiggling with thermal energy, and this thermal agitation creates ripples in the perfectly ordered sea of spins. These collective wave-like excitations of the spin system are called ​​magnons​​.

Each magnon represents one quantum of spin being flipped away from the main alignment, so the more thermal energy you have, the more magnons are excited, and the lower the total saturation magnetization becomes. At low temperatures, this decrease follows a famous relationship known as the ​​Bloch $T^{3/2}$ law​​:

where $M_s(0)$ is the magnetization at absolute zero and $B$ is a constant.

Because the FMR frequency depends directly on $M_s$, it must also depend on temperature! By plugging this temperature-dependent magnetization into our Kittel formula for a thin film, we can predict how the resonance frequency will shift as the material warms up. This provides a beautiful link between the macroscopic, classical-looking phenomenon of resonance and the deep quantum and thermodynamic nature of magnetism. Measuring this frequency shift allows us to probe the population of these ghostly quantum waves—the magnons—sloshing around inside the material.

The simple picture of a precessing top, when combined with the realities of geometry, crystal structure, and thermal fluctuations, blossoms into a rich, predictive, and incredibly useful theory. It is a testament to how a few core principles can illuminate a vast landscape of physical phenomena, from the simplest sphere to the most complex magnetic materials that power our world. And the story doesn't even end here; these same ideas can be extended to understand the coupled dances in more exotic materials like ferrimagnets, showing the remarkable power and reach of this elegant piece of physics.

Now that we have understood the beautiful clockwork of ferromagnetic resonance, let us see what this clock can tell us. It turns out that the Kittel formula is not merely a description of a physical curiosity; it is a master key, unlocking doors to materials science, quantum physics, and the future of computation. By carefully listening to the hum of a precessing magnet, we can learn a spectacular amount about its inner life and even command it to do new and surprising things. The resonance is a window into the soul of the magnet.

Imagine you want to understand the personality of a crystal. Does it have preferred directions? A fundamental trait of magnetic crystals is ​​magnetocrystalline anisotropy​​—an internal energy that makes it easier for the magnetization to point along certain crystallographic axes than others. How can we measure this preference? We can ask the magnet directly, using ferromagnetic resonance (FMR).

If we place a magnetic thin film in a magnetic field and measure its resonance frequency as we rotate the field, we find that the resonance doesn't behave as the simplest Kittel formula predicts. The resonance shifts, and the shifts depend on the orientation. This is because the internal anisotropy fields add to or subtract from the external field, changing the restoring force on the precessing magnetization. By meticulously measuring the resonance field at various angles and fitting the data to a model that includes the anisotropy energy, we can extract the precise numerical values of the anisotropy constants, such as $K_1$ and $K_2$ for a cubic crystal. This is a wonderfully sophisticated procedure that goes far beyond the basic formula, requiring us to calculate the true equilibrium direction of the magnetization and the curvature of the energy landscape at that point. FMR, in this sense, acts as a exquisitely sensitive probe of the material's intrinsic magnetic texture.

We can push this further. How does this magnetic "personality" change with temperature? As a material heats up, thermal fluctuations agitate the atomic spins, causing the overall saturation magnetization $M_s$ to decrease. Does the anisotropy energy change with it? And if so, how? By performing FMR measurements at different temperatures, we can map out the anisotropy constant $K_u(T)$ as a function of temperature. This allows us to test some of the deepest theoretical predictions in magnetism. For instance, the Callen-Callen scaling theory predicts that for many types of anisotropy, the constant should scale with the magnetization as $K_u(T) \propto [M_s(T)]^n$, where the exponent $n$ reveals the microscopic origin of the anisotropy. Using FMR to measure $K_u(T)$ and an independent technique to measure $M_s(T)$, we can experimentally determine this exponent, providing a stringent test of our fundamental understanding of how collective magnetic order emerges from quantum mechanical interactions.

Anisotropy doesn't just come from the crystal lattice, however. Geometry itself is a powerful force. Imagine taking a flat magnetic film and curving it into a cylinder. The very act of introducing curvature creates a new form of magnetic anisotropy. The spins on the curved surface experience an effective magnetic field that depends on the local geometry. This "curvature-induced anisotropy" will, in turn, shift the FMR frequency relative to an identical flat film. By measuring this shift, we can directly quantify the interplay between geometry and magnetism, opening a fascinating field of "curvilinear magnetism" where material properties can be sculpted by simply changing their shape.

A resonance is not defined by its frequency alone. Think of a bell: a well-cast bell rings with a pure tone for a long time, whereas a cracked one gives a dull thud and quickly falls silent. The sharpness of the resonance peak—its linewidth—tells us about how energy is dissipated in the system. In magnetism, this dissipation is elegantly captured by the dimensionless Gilbert damping parameter, $\alpha$. A larger $\alpha$ means more "friction" in the magnetic system, causing the precession to die out more quickly and broadening the FMR peak.

By measuring the FMR linewidth $\Delta H$ at different resonance frequencies $\omega$, we can directly extract the value of $\alpha$. This parameter is of immense practical importance; for applications like magnetic memory, we often want to switch the magnetization quickly, which requires high damping, while for oscillators and filters, we want the resonance to be as sharp as possible, demanding very low damping.

This connection to energy dissipation links FMR to a universal concept in physics: the ​​quality factor​​, or $Q$-factor. For any oscillator, the $Q$-factor is a measure of how many oscillations it undergoes before its energy decays significantly. It is defined as the resonant frequency divided by the linewidth. For an FMR mode, the $Q$-factor is inversely proportional to the Gilbert damping parameter $\alpha$. A low-damping magnet is a high-$Q$ resonator. This beautiful correspondence shows that the precession of trillions of spins in a magnet behaves just like a simple mechanical pendulum or an electrical RLC circuit, revealing a profound unity in the physics of oscillations.

But is this magnetic friction always a simple, constant drag? Nature is often more subtle. If we carefully measure the FMR linewidth as we rotate the magnet, we often find that the linewidth itself changes with angle. This is not typically because the Gilbert damping $\alpha$ is intrinsically anisotropic. Instead, it signals the presence of other, more complex relaxation channels. One of the most important is ​​two-magnon scattering​​. In this process, the uniform precession mode (a magnon with momentum $\mathbf{k}=0$) scatters off of microscopic defects or imperfections in the crystal, creating a pair of new magnons with equal and opposite momentum ($\mathbf{k} \neq \mathbf{0}$). This scattering provides an additional pathway for energy to leak away from the uniform mode, broadening the resonance. The efficiency of this process depends critically on the availability of final states and the nature of the scattering potential, both of which can be highly dependent on the magnetization's orientation. Thus, a detailed map of the FMR linewidth versus angle becomes a powerful diagnostic tool, revealing hidden information about defects and scattering processes within the material.

The principles of FMR are not confined to the physics lab; they are at the heart of revolutionary new technologies. The damping we just discussed, for example, can be more than just an intrinsic property—it can be an engineered feature.

Consider what happens when we place our ferromagnetic film next to a non-magnetic metal. The precessing magnetization in the ferromagnet can pump a current of pure spin angular momentum—not electrical charge—across the interface into the neighboring material. This phenomenon, known as ​​spin pumping​​, acts as an additional source of damping because it carries angular momentum away from the ferromagnetic layer. This increases the measured Gilbert damping parameter. By fabricating two identical samples, one with and one without the adjacent metal layer, and measuring the difference in their FMR linewidths, we can precisely quantify this extra damping. This allows us to determine a crucial parameter for the emerging field of ​​spintronics​​: the spin-mixing conductance $g^{\uparrow\downarrow}$, which measures how transparent the interface is to spin current. This technique is vital for developing new devices like MRAM (Magnetoresistive Random-Access Memory) that use spin currents to write information.

The journey doesn't stop there. It extends all the way to the quantum realm. What happens if our tiny magnet is not in open space, but lives inside a "hall of mirrors"—a high-quality microwave cavity? The resonant mode of the magnet (the Kittel mode, which when quantized is a ​​magnon​​) can couple strongly to the resonant mode of the cavity (a ​​photon​​). When the coupling is strong enough, the magnon and the photon lose their individual identities. They can no longer be considered separate entities. Instead, they form a new hybrid quantum particle, a ​​magnon-polariton​​.

This hybridization manifests as an "avoided crossing" in the system's energy spectrum: as we tune the magnon frequency (by changing the external magnetic field) through the fixed cavity frequency, the two modes repel each other, refusing to become degenerate. The Kittel formula gives us the frequency of the "bare" magnon, which then gets dressed by the interaction with the cavity photon to form the new polaritonic states. This field of ​​cavity magnonics​​ merges magnetism with quantum optics and is a vibrant area of research for quantum information processing and ultra-sensitive detectors.

To conclude our journey, let us step back and appreciate a final, beautiful piece of unification. The very same energy terms that govern the fast dynamics of ferromagnetic resonance—exchange, anisotropy, and magnetostatic (dipolar) energy—also dictate the magnet's static, equilibrium state.

In many materials, a uniformly magnetized state is not the lowest energy configuration. The large magnetostatic energy associated with surface poles can be reduced if the magnet spontaneously breaks up into a complex pattern of ​​magnetic domains​​, with magnetization pointing in different directions in different regions. This, however, comes at the cost of forming domain walls between these regions, which have an energy cost from exchange and anisotropy.

The final, intricate pattern is a result of a delicate compromise, a balancing act between the energy saved by reducing stray fields and the energy spent creating walls. By constructing a simple model that balances the wall energy against the residual magnetostatic energy, one can estimate the characteristic size or period of these domain patterns. This approach, pioneered by Kittel himself, shows that the equilibrium domain period $\Lambda$ scales with the material parameters in a predictable way. This reveals a profound coherence in the physics of magnetism: the same fundamental forces that determine the frequency of a gigahertz oscillation also sculpt the static, micron-scale magnetic landscape. The dance of the spins and the ground they stand on are governed by one and the same set of rules.