🔭 astrophysics

Ultra-slow-roll Inflation with Non-perturbative Non-Gaussianity and Scalar Induced Gravitational Waves

This paper demonstrates that an ultra-slow-roll inflation model with $O(1)$ non-Gaussianity can significantly enhance the abundance of primordial black holes to constitute all dark matter while simultaneously suppressing the associated scalar-induced gravitational wave signals, which nonetheless remain detectable by future observatories like LISA.

Original authors: Manuel Drees, Chenhuan Wang

Published 2026-09-09

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Manuel Drees, Chenhuan Wang

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

In the earliest moments of the universe, a fraction of a second after the Big Bang, space itself underwent a period of explosive expansion known as inflation. This rapid stretching smoothed out the cosmos and planted the seeds for everything we see today, from galaxies to stars. However, this expansion was not perfectly uniform. Tiny quantum fluctuations in the energy field driving inflation created slight ripples in the fabric of space-time. Most of these ripples were gentle, growing into the vast structures of the modern universe, but some theories suggest that under specific conditions, these ripples could have been violent enough to collapse directly into black holes. These hypothetical objects, known as primordial black holes, are distinct from the black holes formed by dying stars; they could have been born in the first instant of existence and might be heavy enough to account for the mysterious dark matter that holds galaxies together.

The question of whether these primordial black holes exist and if they make up all the dark matter has long been a puzzle for cosmologists. A recent study by Manuel Drees and Chenhuan Wang at the University of Bonn offers a fresh look at how such black holes could form. They focused on a specific phase of inflation called "ultra-slow-roll," a period where the field driving the expansion slows down dramatically, almost coming to a halt. This pause allows the tiny quantum ripples to grow much larger than usual. The researchers built a detailed model of this process, tracking the inflation field from the moment it created the patterns we see in the cosmic microwave background radiation all the way to the end of inflation. Their goal was to see if this specific slowdown could generate enough massive fluctuations to create a universe full of asteroid-mass black holes, which are small enough to be invisible to current telescopes but numerous enough to explain the missing mass of the cosmos.

The team discovered that the way the universe transitions out of this slow phase is critical. When the inflation field finally picks up speed again, the change can be smooth or abrupt. The researchers found that if this transition is sharp, it creates a unique statistical effect in the fluctuations. Instead of the ripples following a standard, predictable pattern, they develop a "heavy tail," meaning that extreme, large fluctuations become much more likely than previously thought. This non-standard behavior acts like a multiplier for the formation of black holes. The study shows that even a moderate amount of this unusual statistical behavior is sufficient to boost the number of primordial black holes by orders of magnitude. This means that a model which might have seemed too weak to produce enough dark matter can, with the right transition, easily generate the entire population of dark matter in the form of these tiny black holes.

However, the creation of these black holes comes with a loud side effect. When such massive fluctuations collapse, they inevitably generate ripples in space-time known as gravitational waves. The researchers calculated the strength of these waves, which would be produced at the moment the black holes formed. They found that the same statistical effect that boosts the number of black holes actually weakens the gravitational wave signal for a fixed number of black holes. This is a crucial distinction: because the black holes form so efficiently due to the statistical boost, the universe does not need to generate as many raw fluctuations to reach the same final count. Since gravitational waves are produced by the fluctuations themselves, fewer raw fluctuations mean a quieter signal. Despite this reduction, the team calculated that if these asteroid-mass black holes do indeed make up all the dark matter, the resulting gravitational wave signal would still be strong enough to be detected by future space-based observatories.

The researchers specifically looked at the capabilities of the Laser Interferometer Space Antenna, or LISA, a planned mission designed to listen for gravitational waves from space. Their simulations indicate that the signal from these primordial black holes would be well above the sensitivity threshold of LISA. This creates a powerful test for the theory. If LISA operates and finds no such background hum of gravitational waves, it would effectively rule out the idea that asteroid-mass primordial black holes are the primary source of dark matter in our universe. Conversely, if the signal is detected, it would provide strong evidence for this specific model of inflation and the existence of these ancient black holes. The study also explored how different parameters of the inflation model affect the signal, finding that the shape of the gravitational wave spectrum changes depending on how quickly the inflation field transitions from its slow phase back to normal expansion.

This work provides a complete picture of the inflationary process, connecting the early universe's expansion to the potential existence of dark matter and the gravitational waves we hope to detect soon. By treating the transition between different phases of inflation with care and accounting for the complex, non-linear behavior of the fluctuations, the researchers have shown that the universe could be filled with these tiny black holes without violating any known physical laws. The findings suggest that the search for dark matter is not just a hunt for invisible particles, but also a search for the echoes of the universe's first moments. If the dark matter is indeed made of these primordial black holes, the next generation of gravitational wave detectors will likely hear them, turning a theoretical possibility into an observable reality. The study concludes that while the signal might be quieter than some earlier, simpler models predicted, it remains loud enough to be heard, offering a clear path forward for experimental verification.

Technical Summary: Ultra-slow-roll Inflation with Non-perturbative Non-Gaussianity and Scalar Induced Gravitational Waves

Problem Statement
Primordial Black Holes (PBHs) remain a compelling candidate for Dark Matter (DM), particularly in the asteroid-mass range ( $\sim 10^{-15} M_\odot$ ), where they could constitute the entirety of the DM density. The formation of such PBHs requires primordial curvature perturbations significantly larger than those observed in the Cosmic Microwave Background (CMB). While Ultra-Slow-Roll (USR) inflation can generate the necessary power spectrum enhancement, predicting the resulting PBH abundance is non-trivial. Previous studies often relied on linear approximations between curvature perturbations and density contrast ( $\delta \rho \propto \zeta$ ) and perturbative treatments of non-Gaussianity (NG). However, USR phases can generate $O(1)$ non-Gaussianity, rendering these approximations insufficient. Furthermore, the formation of PBHs via large density fluctuations inevitably generates Scalar Induced Gravitational Waves (SIGW) at second order. A consistent framework is needed to model the transition from CMB scales to the end of inflation, calculate non-perturbative NG effects on PBH abundance, and determine the resulting GW signals for future observatories like LISA.

Methodology
The authors construct a phenomenological inflationary model featuring a USR phase sandwiched between two Slow-Roll (SR) phases.

Inflationary Potential: The inflaton potential $V(\phi)$ is parametrized to describe the entire inflationary history from CMB scales to the end. It consists of a quadratic region (pre-USR), a perfectly flat plateau (USR phase), and a cubic region (post-USR) ensuring a smooth transition back to SR and a valid minimum. The model is defined by free parameters including the USR duration ( $\Delta N$ ), the smoothness of the transition ( $h$ ), and the post-USR duration ( $N_-$ ).
Curvature Perturbations: The Mukhanov-Sasaki equation is solved numerically to obtain the power spectrum $\mathcal{P}_\zeta(k)$ . The evolution of modes during the USR phase is tracked, noting that super-horizon modes are not frozen but can grow exponentially.
Non-Gaussianity ( $\delta N$ Formalism): Instead of using a perturbative expansion for NG, the authors employ the $\delta N$ formalism to derive the exact non-linear relation between the full curvature perturbation $\zeta$ and its Gaussian part $\zeta_G$ . This accounts for the full probability distribution function (PDF) of the fluctuations, which is crucial when $f_{NL} \sim O(1)$ .
PBH Abundance: The calculation avoids the linear approximation $\delta \propto \zeta$ . Instead, it utilizes the compaction function $C$ , which relates non-linearly to $\zeta$ . The PBH abundance is computed by integrating the PDF of the compaction function over the collapse threshold, incorporating critical scaling laws for the final mass.
Gravitational Waves: The SIGW spectrum is calculated from the curvature power spectrum at second order. The authors note that while non-Gaussianity affects the tail of the PDF relevant for PBHs, its impact on the SIGW signal (which integrates over a broader range of modes) is negligible for the $f_{NL}$ values found in this model.

Key Contributions

Non-Perturbative Treatment of NG: The paper demonstrates that $O(1)$ non-Gaussianity, arising naturally from the USR-SR transition, significantly enhances the PBH abundance compared to Gaussian predictions. This enhancement is derived using the full non-linear $\zeta(\zeta_G)$ relation rather than truncated expansions.
Realistic Potential Modeling: Unlike models assuming a sharp cut-off or scale-invariant spectrum post-USR, this work uses a continuous potential that tracks inflation from CMB scales to the end, capturing the specific scale dependence of the power spectrum and the "dip" preceding the peak.
Correlation between PBHs and GWs: The study establishes a quantitative link between the PBH abundance ( $f_{PBH}$ ) and the SIGW signal strength. It reveals a counter-intuitive result: for a fixed PBH abundance, increasing the non-Gaussianity (via a sharper transition parameter $|h|$ ) reduces the required power spectrum amplitude, thereby suppressing the associated GW signal.
Observational Constraints: The authors compute the Signal-to-Noise Ratio (SNR) for future detectors (LISA, BBO, Einstein Telescope) and map the exclusion regions for PBHs as DM based on potential null results.

Results

PBH Abundance: The model successfully produces PBHs that can account for 100% of Dark Matter in the asteroid mass window ( $\sim 10^{-15} M_\odot$ ). The non-Gaussian tail of the PDF enhances the abundance by orders of magnitude compared to Gaussian estimates for the same power spectrum.
GW Signals: The associated SIGW signals peak at frequencies of $O(0.1)$ $O (0.1)$ Hz.
- If PBHs in this mass range contribute a significant component to the total dark matter density, the GW signal is predicted to be well above the sensitivity of LISA, despite the reduction in strength caused by non-Gaussianity.
- The presence of $O(1)$ non-Gaussianity reduces the GW signal strength for a fixed PBH abundance. Specifically, the authors find that $\Omega_{GW} \propto (f_{PBH})^{0.059}$ , meaning the GW signal is relatively insensitive to changes in PBH abundance compared to the abundance's sensitivity to model parameters.
Detectability: Even with the suppression caused by non-Gaussianity, the SNR for LISA and BBO remains $>1$ for scenarios where PBHs form a significant component of DM. Conversely, a null result from LISA would effectively close the window for asteroid-mass PBHs constituting a significant fraction of Dark Matter, unless the mass spectrum is significantly broader than predicted by local features in the potential.
Parameter Dependence: The smoothness parameter $h$ controls the transition speed. Larger $|h|$ (sharper transition) leads to larger non-Gaussianity, steeper drops in the power spectrum at high $k$ , and consequently, a steeper decline in the GW signal at high frequencies.

Significance and Claims
The authors claim that their work provides a more robust and consistent framework for predicting PBH abundance and associated GW signals by treating non-Gaussianity non-perturbatively and using a realistic inflationary potential. They emphasize that while non-Gaussianity facilitates PBH formation, it simultaneously suppresses the GW signal for a fixed DM fraction, a nuance often missed in simpler models.

The paper concludes that future GW observatories, particularly LISA, have the potential to either detect the stochastic background from asteroid-mass PBH formation or rule out this scenario as a significant component of Dark Matter. The authors note that their specific treatment of the potential and non-Gaussianity leads to lower GW signals than previous studies assuming scale-invariant spectra, making the detection challenge more stringent but still within the projected reach of LISA if PBHs contribute significantly to the DM. They acknowledge caveats regarding loop corrections and the precise profile of curvature perturbations but argue that their main qualitative conclusions regarding the detectability and the inverse relationship between NG-enhanced PBH abundance and GW signal strength remain robust.

More like this