Ferromagnetic resonance modes in trilayer artificial spin ices subject to interfacial Dzyaloshinskii-Moriya interaction

This study numerically demonstrates that interfacial Dzyaloshinskii-Moriya interaction induces non-reciprocity in strongly coupled trilayer square artificial spin ices, leading to frequency-split edge modes whose interference patterns are governed by the interplay between DMI magnitude, external magnetic fields, and stray fields.

The Gist

Here is an explanation of the paper, translated into everyday language with some creative analogies.

The Big Picture: Building a 3D Magnetic Lego Set

Imagine you are building a city out of tiny, magnetic Lego bricks. In the world of physics, these bricks are called Artificial Spin Ices (ASIs). Usually, scientists build these cities in a flat, 2D layer. But recently, they started stacking them up to make 3D structures, like a sandwich with three layers of magnetic bricks.

This paper is about what happens when you take that 3D magnetic sandwich and add a special "twist" to it. The researchers wanted to see how this twist changes the way energy (specifically, magnetic waves) travels through the structure.

The Ingredients: The Magnetic Sandwich

The researchers built a specific type of sandwich:

1. The Top Layer: Made of Permalloy (Py), a soft, squishy magnetic material. Think of this like a piece of jelly. It's easy to move and change shape.
2. The Bottom Layer: Made of CoFe, a hard, stiff magnetic material. Think of this like a rock. It's stubborn and doesn't want to move easily.
3. The Filling: A tiny gap (just 5 nanometers thick) separating the jelly and the rock.

Because the gap is so small, the "jelly" and the "rock" are constantly talking to each other. The rock's magnetic field pushes on the jelly, and the jelly pushes back. They are strongly coupled, meaning they act almost like a single unit rather than two separate layers.

The Secret Ingredient: The "Dzyaloshinskii-Moriya Interaction" (DMI)

Now, imagine you have a rule for how these magnetic bricks behave. Usually, if you push a wave from left to right, it behaves the same as pushing it from right to left. This is called reciprocity.

But the researchers introduced a special ingredient called DMI. You can think of DMI as a one-way street sign or a handedness (like being left-handed vs. right-handed).

• When DMI is present, the magnetic waves stop being fair. They behave differently depending on which direction they travel.
• If a wave tries to go "forward," it might speed up. If it tries to go "backward," it might slow down or change its shape entirely.

The Experiment: Tuning the Radio

The researchers used a computer to simulate this magnetic sandwich. They did two main things:

1. Checking the Connection: First, they made sure the "jelly" and "rock" layers were close enough to talk to each other. They found that at a 5nm gap, they were best friends, influencing each other's magnetic states heavily.
2. Listening to the Music (Resonance): They then "shook" the system with a magnetic pulse (like plucking a guitar string) and listened to the frequencies it produced. This is called Ferromagnetic Resonance (FMR).

The Surprising Discoveries

Here is what they found when they turned on the "one-way street" (DMI):

1. New "Edge" Songs Appear
In a normal magnetic system, waves usually bounce around in the middle of the brick (like a standing wave in a room). But with the DMI twist, new waves started appearing specifically at the edges of the bricks.

• Analogy: Imagine a guitar string. Usually, the whole string vibrates. But with this new "twist," the vibration suddenly concentrates only at the very ends of the string, creating a new, unique sound.

2. The "Interference" Dance
The researchers found that the direction of the DMI (positive or negative) acted like a switch.

• Sometimes, the waves from the top layer and the bottom layer would clash (destructive interference), canceling each other out and making the signal disappear.
• Other times, they would team up (constructive interference), making the signal much louder and stronger.
• Analogy: Think of two people clapping. If they clap at the exact same time, it's loud (constructive). If one claps while the other is silent, or they clap out of sync, the sound is weak or weird (destructive). The DMI and the external magnetic field controlled who was clapping when.

3. The "Soft" Layer Does the Heavy Lifting
Interestingly, the "rock" layer (CoFe) barely changed its behavior. It was too stiff to be affected by the DMI. The "jelly" layer (Permalloy), however, was very sensitive. It was the jelly layer that started singing these new "edge songs" and doing the interference dance.

• Takeaway: To control these 3D magnetic structures, you only really need to tweak the soft layer.

Why Does This Matter?

This research is a big step forward for Magnonics—a field that wants to use magnetic waves (magnons) instead of electricity to carry information in computers.

• Reconfigurable Circuits: Because the waves change based on the magnetic field and the DMI, we could potentially build computer chips that can be "rewired" on the fly just by changing the magnetic settings.
• Topological Protection: The "edge modes" found here are special. In physics, edge states are often "protected," meaning they are very hard to break or scatter. This could lead to super-stable data storage or transmission that doesn't lose information easily.
• 3D Computing: By proving that we can control waves in a 3D stack, this opens the door to building much denser, more complex magnetic computers that aren't limited to flat surfaces.

Summary

The paper shows that by stacking two different magnetic materials and adding a "one-way street" rule (DMI), we can create new types of magnetic waves that live on the edges of the material. These waves can be turned on, off, or amplified by simply adjusting the magnetic field, offering a powerful new toolkit for building the next generation of magnetic computers.

Technical Summary

Here is a detailed technical summary of the paper "Ferromagnetic resonance modes in trilayer artificial spin ices subject to interfacial Dzyaloshinskii-Moriya interaction."

1. Problem Statement

Artificial Spin Ices (ASIs) are metamaterials composed of arrays of nanomagnets that exhibit reconfigurable magnetic states. While 2D ASIs have been studied extensively, their utility as magnonic crystals (materials that control spin waves) is limited by weak dynamic coupling between elements, which hinders the formation of robust magnonic band structures.

• The Gap: Existing 2D ASIs rely on stray fields for static coupling, but dynamic coupling (necessary for band structures) is often too weak.
• The Challenge: Creating 3D ASIs (e.g., trilayers) increases coupling but introduces complexity in understanding how interfacial effects, specifically the Dzyaloshinskii-Moriya Interaction (DMI), influence ferromagnetic resonance (FMR) modes and spin wave non-reciprocity in confined geometries.
• Objective: To investigate how interfacial DMI affects the static magnetization states and dynamic FMR modes in a strongly coupled trilayer square ASI, specifically looking for non-reciprocal effects and mode splitting.

2. Methodology

The authors employed micromagnetic simulations using the software MuMax3 to model the system.

• System Geometry:
  • Trilayer Element: A stadium-shaped nanoisland ($L=200$ nm, $W=100$ nm, $T=5$ nm) composed of a top Permalloy (Py) layer and a bottom CoFe layer, separated by a non-magnetic spacer of thickness $G$.
  • ASI Array: A square lattice of these trilayers with a center-to-center distance $d = 300$ nm.
  • Materials:
    • Py (Permalloy): $M_s = 790$ kA/m, $A = 10$ pJ/m, $\alpha = 0.01$.
    • CoFe: $M_s = 1700$ kA/m, $A = 26$ pJ/m, $\alpha = 0.0004$.
• Coupling Optimization:
  • Hysteresis loops and FMR spectra were simulated for varying spacer gaps ($G = 5, 20, 40$ nm).
  • Result: $G = 5$ nm was selected as it ensures strong static and dynamic coupling (evidenced by complex hysteresis features and multiple coupled FMR peaks), whereas $G=40$ nm resulted in decoupled layers.
• DMI Implementation:
  • Interfacial DMI was introduced in both layers.
  • Two configurations were tested: DMI+ (same sign for Py and CoFe) and DMI- (opposite signs).
  • DMI strength ($D$) was swept from $0$to$\pm 1$ mJ/m².
• Simulation Protocol:
  • Initial state: Antiparallel remanent state.
  • Excitation: A 100 mT field pulse (10 ps) normal to the plane.
  • Analysis: Time-resolved magnetization recorded for 40 ns to generate FMR spectra (FFT of $m_z$) and mode profiles.

3. Key Contributions

1. Identification of Strong Coupling Regime: Established that a 5 nm gap in a Py/CoFe trilayer creates a regime where static and dynamic couplings are sufficient to treat the system as a unified magnonic entity rather than independent layers.
2. DMI-Induced Mode Splitting: Demonstrated that interfacial DMI, in conjunction with the ASI's internal stray fields, induces frequency splitting of edge modes. This splitting is dependent on the sign and magnitude of the DMI.
3. Interference Mechanisms: Revealed that the interplay between DMI and external magnetic fields leads to constructive and destructive interference of spin wave modes. This results in the selective enhancement or suppression of specific modes (e.g., bulk vs. edge modes) depending on the DMI polarity.
4. Layer-Specific Sensitivity: Found that the "soft" Py layer is highly sensitive to DMI and dictates the formation of new standing wave modes, while the "hard" CoFe layer remains relatively insensitive, acting primarily as a bias field provider.

4. Key Results

• Static Properties (Hysteresis):
  • At $G=5$ nm, the system exhibits strong coupling. The CoFe stray field opposes the external field, causing premature reversal in the Py layer and complex hysteresis loops that deviate from the standard Stoner-Wohlfarth model.
• Dynamic Properties (FMR Spectra):
  • Mode Identification: Four primary modes were identified:
    • B (Bulk): Zeroth-order mode.
    • E (Edge): Localized at the nanoisland edges.
    • M1 & M2: Standing wave modes with distinct DMI dependence.
  • DMI Dependence:
    • M1: Dominant for negative DMI; exhibits a strong red-shift as DMI increases.
    • M2: Dominant for positive DMI; frequency varies slightly.
    • Bulk Mode Splitting: The bulk mode splits at high DMI ($\approx 1$ mJ/m²), indicating a change in the equilibrium magnetization state.
  • Field Dependence & Interference:
    • Under an external field, modes split further.
    • Destructive Interference: For $D = 1$ mJ/m², the bulk mode at remanence has vanishing amplitude due to destructive interference between layers.
    • Constructive Interference: For $D = -1$ mJ/m², the bulk mode has a large amplitude.
    • Edge Mode Splitting: In the $D = -1$ mJ/m² case, the edge mode ($E2$) splits into several weaker modes, whereas in the $D = 1$ mJ/m² case, a well-defined mode appears at $\approx 10$ GHz.
• Mode Profiles:
  • M1 Mode: Primarily a transverse standing wave localized in the Py layer.
  • M2 Mode: A longitudinal standing wave (combination of edge and bulk).
  • Asymmetry: The Py layer shows significant "buckling" of magnetization at edges due to low $M_s$ and CoFe stray fields, enabling strong oscillation. The CoFe layer maintains a more uniform state.

5. Significance and Implications

• Tunable Magnonic Band Structures: The study proves that DMI can be used as a tuning knob to manipulate the magnonic band structure in 3D ASIs, enabling reconfigurable bandgaps and frequency combs.
• Non-Reciprocity in Nanostructures: It confirms that non-reciprocity induced by DMI in small nanoislands can significantly alter long-range states stabilized in ASIs, a crucial factor for designing directional spin wave devices.
• Topological Protection: The findings suggest a pathway toward creating topologically protected magnons in 3D spin ices, particularly when interfacing soft magnetic layers with heavy metals to induce DMI.
• Experimental Guidance: The results indicate that for experimental realization, only the soft magnetic layer (Py) needs to be interfaced with a heavy metal to observe the predicted FMR behaviors, simplifying fabrication requirements.

In conclusion, this work bridges the gap between 2D ASIs and complex 3D geometries, demonstrating that interfacial DMI is a powerful tool for engineering spin wave dynamics, mode interference, and non-reciprocal transport in artificial spin ice metamaterials.