⚛️ phenomenology

Axion Isocurvature Perturbations Survive the Scaling Evolution of Axion Domain Walls

This paper demonstrates that sizable axion isocurvature perturbations can survive the scaling evolution and eventual annihilation of axion domain walls, as the biased vacuum energy released during collapse inherits and transfers superhorizon inflationary correlations to the axion dark matter density.

Original authors: Naoya Kitajima, Junseok Lee, Kai Murai, Fuminobu Takahashi, Wen Yin

Published 2026-08-17

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Naoya Kitajima, Junseok Lee, Kai Murai, Fuminobu Takahashi, Wen Yin

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Cosmic Echo: Why the Universe Remembers Its First Moments

Imagine the universe as a giant, expanding balloon. Long ago, when this balloon was tiny and incredibly hot, it underwent a rapid stretching phase called "inflation." During this time, the fabric of space was so smooth that tiny quantum jitters—like the faintest ripples on a pond—were stretched out to become massive, cosmic-sized waves. These ripples are the seeds of everything we see today: galaxies, stars, and even the invisible "dark matter" that holds galaxies together.

One of the leading candidates for this mysterious dark matter is a ghostly particle called the axion. Think of axions as a field of invisible energy that fills the universe, like a thick fog. In the early universe, this fog was calm, but as the cosmos cooled, the axion field had to "choose" a resting spot, much like a ball rolling down a hill to settle in a valley. If the hill has many identical valleys, the ball might pick different ones in different parts of the universe. This choice creates boundaries called domain walls, where the field suddenly switches from one valley to another. These walls are like invisible fences separating different regions of the cosmos.

Scientists have long wondered: Do these cosmic fences erase the memory of the original ripples from inflation? The standard idea was that as the universe expanded and these walls moved around, they would eventually smooth everything out, wiping the slate clean and leaving no trace of the initial cosmic jitters. If this were true, it would be great news for the axion theory, as it would mean we don't have to worry about the "noise" from the early universe messing up our current observations. But what if the walls don't forget? What if the universe is like a stubborn old record player that keeps playing the same scratchy song no matter how much you try to clean the needle?

The Paper's Discovery: The Walls Don't Forget

In this paper, a team of physicists revisits the story of these axion domain walls, specifically looking at a scenario where the initial "jitters" from inflation were huge—so huge that the axion field didn't just pick one valley, but scattered across many different valleys right from the start. They wanted to see if this chaotic start would help the walls forget their origins, or if the memory would survive.

Using powerful computer simulations and mathematical arguments, the authors found that the "forgetting" theory is likely wrong. Even when the domain walls enter a chaotic, scaling phase where they look like they are losing their structure, they actually retain the super-long-distance correlations from the very beginning.

Here is the core of their finding, explained with a simple analogy:

Imagine a giant, bouncy trampoline (the universe) covered in thousands of small, colorful balls (the axion field). In the beginning, the trampoline is shaken violently (inflation), sending the balls bouncing into different corners. Some corners get more balls than others. Now, imagine that the balls start sticking together to form walls.

The old idea was that as these walls moved and crashed into each other, they would mix the balls up so thoroughly that you couldn't tell which corner they started in. It would be like stirring a cup of coffee until the sugar is perfectly dissolved; you can't taste the sugar anymore.

However, this paper shows that the "stirring" isn't perfect. When the walls finally collapse and disappear, they release a burst of energy. The authors discovered that this burst of energy remembers exactly where the balls started. If a large region of the trampoline started with a "heavy" concentration of balls, the energy released when the walls in that region collapsed was also "heavy." If another region was "light," the energy released there was "light."

The authors ran numerical lattice simulations (essentially a high-tech video game of the universe) to prove this. They watched the walls form, move, and collapse. They saw that even after the walls were gone, the leftover energy (which becomes dark matter) still had a "fingerprint" of the original inflationary ripples. The large-scale patterns of the early universe were transferred directly to the dark matter, surviving the entire chaotic process.

Why This Matters and What It Rules Out

This finding is a significant challenge to a recent idea that tried to save the axion theory. Some scientists had suggested that if the axion field populated many different valleys (vacua) at the start, the "noise" or isocurvature perturbations (which are like unwanted static on a radio signal) would average out and disappear. They thought this would make the axion a perfect, quiet candidate for dark matter.

The paper argues against this. They show that while the "misalignment" part of the axion (the balls just sitting in the valleys) might get smoothed out, the part coming from the collapse of the domain walls does not. The energy released during the collapse inherits the "static" from the inflationary fluctuations.

The authors are quite sure about this based on their simulations. They found that the "static" (isocurvature perturbations) remains sizable and nearly scale-invariant (meaning it looks the same at all sizes), just like the original inflationary ripples. This means that simply populating many vacua is not enough to hide the axion from our observations.

In fact, the paper suggests that if the axion dark matter was created this way, it would create a "loud" signal that we should be able to detect. The authors calculated that for this scenario to work without breaking the rules set by the Cosmic Microwave Background (the afterglow of the Big Bang), the universe would have had to be incredibly hot—hotter than 10^16 GeV. This is a temperature so high that it likely exceeds the maximum heat the universe could have reached after inflation, given what we know about the energy of the Big Bang.

The Bottom Line

So, what does this mean for the curious teenager? It means the universe is a better historian than we thought. Even when cosmic structures like domain walls go through a chaotic "scaling" phase where they seem to lose their shape, they don't actually erase the memory of how they were born. The "ghost" of the inflationary fluctuations survives the collapse of the walls and gets imprinted onto the dark matter.

The paper concludes that trying to hide the axion's "noise" by just filling up many valleys is a dead end. The noise is still there, loud and clear, carried by the energy of the collapsing walls. Unless there is some other trick—like the axion's properties changing over time to turn the "noise" into a high-pitched squeal that we can't hear yet—the standard axion model with these huge initial fluctuations is likely ruled out by current observations. The universe remembers, and it's telling us that the axion story might need a rewrite.

Technical Summary: Axion Isocurvature Perturbations Survive the Scaling Evolution of Axion Domain Walls

Problem Statement
The paper addresses the cosmological evolution of axion domain walls seeded by inflationary fluctuations, specifically challenging the conventional assumption that the "scaling regime" of topological defect networks erases memory of initial conditions. In standard scenarios involving thermal or white-noise initial conditions (Kibble mechanism), defect networks are expected to lose superhorizon correlations as they approach a scaling solution where the characteristic separation is of order the Hubble radius. However, for axions with large inflationary fluctuations ( $H_{\text{inf}} \gg f_\phi$ ), different patches of the universe populate multiple minima of the axion potential before oscillation begins.

A recent proposal suggested that populating many vacua could suppress the long-wavelength isocurvature perturbations of axion dark matter by averaging over the periodic potential, thereby evading cosmic microwave background (CMB) constraints. This paper investigates whether this suppression mechanism holds when the axion abundance is dominated by the collapse of domain walls rather than the conventional misalignment mechanism. The central question is whether the superhorizon correlations inherent in inflationary fluctuations survive the scaling evolution and subsequent annihilation of the domain-wall network.

Methodology
The authors employ a combination of analytical arguments and numerical lattice simulations to study the dynamics of axion domain walls.

Analytical Framework:
- The authors model the axion potential $V(\phi)$ with a periodic term and a small bias term $V_b(\phi)$ that lifts the degeneracy of vacua, driving the walls to collapse.
- They classify the evolution into two regimes based on the magnitude of inflationary fluctuations relative to the bias period: the Single-Bias-Period Regime (fluctuations span multiple minima of $V(\phi)$ but less than one period of the bias) and the Multiple-Bias-Period Regime (fluctuations span multiple periods of the bias).
- They derive the power spectrum of the axion energy density by treating the domain-wall annihilation as a local process. They argue that the released biased vacuum energy inherits the superhorizon correlations of the initial field fluctuations. The coarse-grained vacuum fractions within a Hubble volume are modulated by long-wavelength modes of the axion field, leading to a transfer of these correlations to the axion density.
Numerical Simulations:
- The authors perform two-dimensional lattice simulations in a radiation-dominated universe.
- Initial conditions are set with a scale-invariant power spectrum for the axion field fluctuations, ensuring that correlations exist on superhorizon scales.
- Simulations are run both without bias (to confirm scaling behavior) and with a small quadratic bias to induce wall collapse.
- They track the evolution of the domain-wall network, the vacuum histogram, and the power spectrum of the axion field and energy density.

Key Contributions and Results

Survival of Superhorizon Correlations: The primary finding is that entering the scaling regime does not erase primordial superhorizon correlations. Even when the domain-wall network reaches a scaling solution and subsequently collapses due to a potential bias, the resulting axion energy density retains the large-scale fluctuations of the initial inflationary field.
Mechanism of Transfer: The authors demonstrate that the annihilation process acts as a local transfer function. The energy released during the collapse of biased domain walls depends on the local vacuum fractions, which are determined by the long-wavelength axion field value. Consequently, the biased vacuum energy inherits the nearly scale-invariant spectrum of the inflationary fluctuations.
Failure of the "Many Vacua" Suppression: While the conventional misalignment contribution to axion abundance can be suppressed by averaging over many vacua, the contribution from domain-wall collapse is not. The paper shows that sizable isocurvature perturbations are generated by axions produced during wall annihilation, even when many minima are populated.
Numerical Verification:
- Simulations confirm that domain walls formed from scale-invariant fluctuations exhibit robust scaling behavior and that large domains extend across multiple Hubble patches.
- In the presence of a bias, the network collapses, but the spatial distribution of the resulting axion energy density correlates with the initial vacuum distribution.
- The power spectrum of the axion density fluctuations shows a characteristic peak at the Hubble scale (from the collapse dynamics) and a significant low- $k$ tail that increases with the mean field displacement, consistent with the analytical prediction of a nearly scale-invariant component.
Isocurvature Constraints: The authors derive generic constraints on dark matter produced by domain-wall collapse. They find that for a scale-invariant spectrum (or one with a small tilt), the required formation temperature to satisfy CMB isocurvature bounds exceeds $10^{16}$ GeV. This is generally incompatible with the maximum temperature achievable after inflation, implying that such a dark matter scenario is excluded by current observations unless the spectrum is significantly blue-tilted.
Composite Walls: In the multiple-bias-period regime, the authors discuss the formation of composite domain walls (sandwiches of elementary walls) which can be long-lived if the true vacua are degenerate, potentially leading to a domain-wall problem or distinct gravitational wave signatures.

Significance and Claims
The paper claims to clarify a misconception in the literature regarding the "loss of memory" in scaling defect networks. It asserts that for inflationary initial conditions, causality dictates that superhorizon correlations cannot be erased by local defect dynamics.

The authors conclude that the mechanism proposed in recent literature to evade axion isocurvature constraints by populating multiple vacua is generically insufficient. The energy released during domain-wall annihilation preserves the superhorizon correlations of the inflationary fluctuations, transferring them to the axion energy density. Therefore, unless the inflationary fluctuations are modified (e.g., via a time-dependent decay constant to induce a blue tilt) or the domain-wall contribution to dark matter is subdominant, the standard axion domain-wall scenario remains in tension with CMB isocurvature constraints.

The paper also highlights that the specific spectral index of the isocurvature perturbation (controlled by the tilt of the initial fluctuations or the bias structure) is crucial for determining the viability of the model against observational bounds, with small-scale probes (like Lyman- $\alpha$ or 21 cm) becoming relevant for non-standard spectral indices.

The Cosmic Echo: Why the Universe Remembers Its First Moments

The Paper's Discovery: The Walls Don't Forget

Why This Matters and What It Rules Out

The Bottom Line

Technical Summary: Axion Isocurvature Perturbations Survive the Scaling Evolution of Axion Domain Walls

More like this