Revealing domain wall stability during ultrafast demagnetization

Using ultrafast sub-wavelength extreme ultraviolet imaging, researchers demonstrated that magnetic domain walls in ferro- and ferrimagnetic thin films remain remarkably stable in position, shape, and width even during up to 50% ultrafast demagnetization, revealing the localized nature of photoinduced demagnetization and offering new insights for all-optical magnetic control.

The Gist

Imagine a city made entirely of tiny magnets. In this city, the "streets" are called domain walls. These are the boundaries where the magnetic direction flips, like a street where houses on the left face North and houses on the right face South.

For years, scientists have been trying to understand what happens to these "streets" when you hit the city with a massive, ultra-fast laser pulse (like a sudden, intense heatwave). The big question was: Does the laser melt the streets, making them wider and shifting their location? Or do they stay put?

Some previous experiments suggested the streets were chaotic, moving at incredible speeds and getting wider. But those experiments were like trying to watch a hummingbird's wings with a blurry, slow camera—you could see something moving, but you couldn't be sure what it was.

This paper introduces a brand-new, super-powered camera that acts like a high-speed, super-magnifying glass. It can see things that are smaller than the wavelength of light itself (sub-wavelength) and capture them in less than a blink of an eye (femtoseconds).

Here is what they discovered, using some simple analogies:

1. The "Ghost" in the Machine

When the researchers hit their magnetic films with a laser, they expected the "streets" (domain walls) to get messy. They thought the heat would make the walls wobble, stretch out, or slide around like a rubber band snapping.

The Surprise: The streets didn't budge.
Even when the laser removed up to 50% of the magnetism (imagine turning off half the lights in the city), the boundaries between the North-facing and South-facing houses remained exactly where they were, exactly the same width, and exactly the same shape.

It's as if you poured a bucket of boiling water over a sandcastle, but the moat and the walls stayed perfectly intact, only the water inside the castle evaporated.

2. The "Rigid Fence" vs. The "Wobbly Rope"

Think of the magnetic domain wall as a fence.

• Old Theory: Scientists thought the fence was made of a wobbly rope. When you heated it, the rope would get loose, stretch out, and swing wildly.
• New Discovery: The fence is actually made of steel. It is incredibly rigid. The laser energy was absorbed by the "people" inside the houses (the electrons and spins), but it didn't shake the fence itself. The fence only starts to break and move if you hit it with enough energy to completely destroy the neighborhood (irreversible switching).

3. Why Did We Get It Wrong Before?

The paper explains that previous experiments were like trying to guess the shape of a moving car by looking at its blurry shadow in the fog.

• The Blur: Old methods had to guess the shape of the walls based on statistical averages, which led to the idea that the walls were moving super fast.
• The Clarity: This new technique takes a direct, crystal-clear photo. It turns out the "fast movement" seen before was likely an illusion caused by the walls breaking apart in specific spots, not the whole wall sliding across the city.

4. The "Tipping Point"

The researchers found a "tipping point."

• Below 50% demagnetization: The magnetic walls are rock-solid. They don't move, they don't stretch. This is great news for future technology because it means we can control magnets very quickly without accidentally messing up the data storage.
• Above 50% demagnetization: If you hit it too hard, the "streets" do start to break. Random little islands of magnetism flip over, like dominoes falling in a chaotic pattern. This is when the system becomes unstable.

Why Does This Matter?

This discovery is a game-changer for spintronics (the next generation of super-fast, super-dense computer memory).

• The Good News: Because the walls are so stable, we can use lasers to switch data on and off incredibly fast without worrying about the "roads" getting blurry or moving to the wrong place.
• The Challenge: We need to be careful not to hit the system too hard, or we might accidentally scramble the data.

In a nutshell:
Scientists used a super-camera to watch magnetic "streets" during a laser storm. They found that, contrary to what they thought, the streets are surprisingly tough and don't move or stretch until the storm is absolutely catastrophic. This gives us a clear roadmap for building faster, more reliable computers that use light instead of electricity to store data.

Technical Summary

1. Problem Statement

The manipulation of nanoscale spin textures (such as magnetic domain walls and skyrmions) using femtosecond laser pulses is crucial for next-generation high-speed, high-density spintronics. However, understanding the underlying mechanisms of photoinduced demagnetization and domain wall (DW) dynamics has been hindered by a lack of imaging techniques that simultaneously offer:

• Nanometer spatial resolution: Required to resolve DW widths (typically 10–20 nm).
• Femtosecond temporal resolution: Required to capture dynamics occurring on sub-picosecond timescales.

Previous studies faced a trade-off:

• Scattering experiments (X-ray/XUV): Offered femtosecond resolution but relied on reciprocal space data, requiring assumptions about domain morphology and failing to provide direct real-space images. These studies often inferred ultrafast DW velocities ($\sim$70 km/s) and significant broadening ($\sim$50%).
• Microscopy (Optical/Soft X-ray): Provided real-space imaging but lacked either the spatial resolution (optical) or the temporal resolution (soft X-ray synchrotrons, limited to tens of picoseconds) to track ultrafast DW dynamics directly.

2. Methodology

The authors developed a novel ultrafast sub-wavelength imaging technique using a table-top High-Harmonic Generation (HHG) source to overcome these limitations.

• Experimental Setup:

  • Pump-Probe Scheme: A femtosecond Ti:sapphire laser (800 nm) serves as the pump to excite the sample. A time-delayed, circularly polarized Extreme Ultraviolet (XUV) pulse (20.8 nm, $\sim$60 eV) acts as the probe, tuned to the Cobalt $M_{2,3}$ absorption edge to maximize Magnetic Circular Dichroism (XMCD) contrast.
  • Sample: Thin films of ferrimagnetic $Tb_{26}Co_{74}$ and ferromagnetic Co/Pd multilayers with strong perpendicular magnetic anisotropy.
  • Imaging Geometry: The sample is partially covered by an opaque gold mask with a 3-$\mu$m diameter circular aperture. The XUV beam illuminates this aperture, and the far-field diffraction pattern is recorded by a CCD camera.
  • Reconstruction: Iterative phase retrieval algorithms (RAAR and Error Reduction) are used to reconstruct the real-space electric field amplitude and phase from the diffraction data. By taking the phase difference between left- and right-circularly polarized probes ($\phi_L - \phi_R$), the magnetic signal is isolated from non-magnetic contributions.

• Quantitative Analysis:

  • Resolution: Achieved a spatial resolution of 13.5 nm (sub-wavelength) and a temporal resolution of < 40 fs.
  • DW Extraction: Local DW profiles were extracted by fitting the phase contrast data to a hyperbolic tangent function: $M_z(x) = M_0 \tanh((x-x_0)/w)$, where $x_0$ is the position and $w$ is the width.
  • Precision Validation: The precision of the measurements was validated by comparing independent datasets and performing Monte Carlo simulations to account for signal-to-noise ratio (SNR) limitations.

3. Key Contributions

1. Technological Advancement: Demonstrated the first real-space imaging of magnetic domain walls with simultaneous sub-10 nm spatial resolution and femtosecond temporal resolution, bridging the gap between scattering and microscopy techniques.
2. Quantitative Precision: Established a measurement precision for DW width and position of < 1 nm (specifically $\sigma_{\delta w} \approx 0.8$ nm for width and $\sim$1.5 nm for position baseline), allowing for the detection of minute changes previously obscured by noise or lower resolution.
3. Direct Observation of Stability: Provided direct evidence that, contrary to some scattering-based inferences, domain walls remain remarkably stable during the initial stages of ultrafast demagnetization.

4. Key Results

• Domain Wall Invariance:

  • For demagnetization levels up to $\sim$50–60%, the position ($x_0$), shape, and width ($w$) of the domain walls remain invariant within the experimental precision ($\sim$1 nm).
  • The spatially averaged DW width ($w_{avg}$) showed no statistically significant increase (change $< 0.3$ nm) even at high pump fluences inducing $\sim$40% magnetization suppression.
  • No systematic displacement of DWs was observed; apparent movements were attributed to fitting uncertainties caused by reduced XMCD contrast in the demagnetized state.

• Threshold for Instability:

  • Stochastic Switching: Only when the pump fluence induced demagnetization exceeding $\sim$50% did the researchers observe irreversible, stochastic nanoscale domain switching. This suggests a threshold behavior where the domain structure becomes unstable only under extreme excitation.
  • Irreversible Changes: At high fluences ($\sim$9 mJ/cm$^2$), local regions exhibited magnetization switching, leading to permanent changes in the domain pattern.

• Comparison with Previous Models:

  • The results contradict previous scattering studies that inferred DW velocities of $\sim$70 km/s and significant broadening. The authors suggest those observations may have been misinterpreted signatures of irreversible structural changes or stochastic switching rather than reversible ultrafast DW motion.
  • The findings support a localized nature of photoinduced demagnetization, where spin transport effects are limited and do not cause immediate, large-scale DW deformation.

5. Significance

• Fundamental Physics: The study resolves a long-standing debate regarding the mechanism of ultrafast demagnetization. It establishes that the primary response of magnetic textures to femtosecond laser pulses is a reduction in magnetization magnitude rather than immediate structural deformation or rapid motion of the walls.
• Spintronics Applications: The robustness of domain walls under intense laser excitation is critical for the development of all-optical magnetic switching (AOS) devices. It suggests that stable, ultra-fast toggling of magnetic bits is possible without destroying the domain wall integrity, provided the excitation is controlled below the stochastic switching threshold.
• Methodological Impact: The demonstrated XUV microscopy technique provides a powerful, generalizable tool for probing nanoscale dynamics in spintronic materials, complex magnetic systems (including near compensation points), and even non-magnetic chemical/structural dynamics with unprecedented spatiotemporal precision.

In conclusion, this work redefines the upper bounds for reversible ultrafast domain wall dynamics, proving that domain walls are robust entities that maintain their structural integrity during the initial phase of laser-induced demagnetization, challenging previous models of ultrafast spin transport and motion.