Impact of the axion-like self-interactions in gravitational atoms for LISA

This paper demonstrates that the Laser Interferometer Space Antenna (LISA) can detect and constrain the masses and self-interaction strengths of axion-like particles forming gravitational atoms around black holes by analyzing the dephasing effects they induce on gravitational waveforms from extreme- and intermediate-mass-ratio inspirals.

The Gist

Imagine the universe is filled with a ghostly, invisible fog made of ultra-light particles called axions. These particles are so light and numerous that they don't just float around; they can clump together, forming massive, invisible clouds around heavy objects like black holes. The authors of this paper call these clouds "gravitational atoms."

Just like a real atom has a nucleus (the black hole) and an electron cloud (the axion fog), these "gravitational atoms" have a structure, but on a cosmic scale.

Here is what the paper explores, broken down into simple concepts:

1. The Setup: A Cosmic Dance

Imagine two black holes orbiting each other. One is a giant (the primary), and the other is a smaller companion. They are spiraling inward, getting closer and closer until they eventually crash together. This dance emits ripples in space-time called gravitational waves, which we can detect with instruments like LISA (a future space-based detector).

Usually, we expect this dance to follow a very specific rhythm based on gravity alone. However, if the giant black hole is surrounded by one of these "gravitational atom" clouds, the rhythm changes. The smaller black hole has to swim through this thick fog as it orbits.

2. The Mechanism: How the Cloud Grows

Previous ideas suggested these clouds formed because the black hole was spinning fast and "sucking" energy out of the fog (like a whirlpool).

This paper proposes a different, more direct way the cloud forms: Self-Interaction.
Think of the axion particles as people at a party who really like to hug each other. Because they have a "self-interaction" (they attract one another), they naturally clump together around the black hole over time. The paper uses a new model to calculate exactly how fast this cloud grows and how dense it gets, starting from a very thin background fog in the galaxy.

3. The Effect: The "Drag" on the Dance

As the smaller black hole orbits through this axion cloud, two main things happen:

• Dynamical Friction (The Drag): Imagine running through a pool of water versus running through air. The water slows you down. The axion cloud acts like the water. As the small black hole moves, it pulls the axions along with it, creating a wake. This drag steals energy from the orbit, causing the two black holes to spiral together faster than they would in empty space.
• Accretion (The Snack): The small black hole also eats some of the axion particles, gaining a tiny bit of mass, though the paper finds this effect is much smaller than the drag.

4. The Result: A Different Song

Because of this drag, the gravitational waves emitted by the black holes change.

• The Phase Shift: In music, if you play a song slightly out of time, it sounds "off." In gravitational waves, this is called dephasing. The cloud causes the black holes to get out of sync with the "vacuum" rhythm (the rhythm they would have in empty space).
• The Signature: This isn't just a small glitch; it's a distinct pattern. The paper calculates that for certain sizes of black holes and certain types of axions, this "off-key" signal is loud enough for LISA to hear.

5. What LISA Can "See"

The authors ran simulations to see what LISA could detect. They found that:

• The Sweet Spot: There is a specific "Goldilocks zone" for the mass of the axion particles. If they are too heavy, the cloud is too small to matter. If they are too light, the cloud is too spread out to create drag. But in the middle range, the effect is strong.
• The Measurement: If LISA detects a signal with a high enough "signal-to-noise ratio" (a clear, loud signal), it can distinguish between a black hole in empty space and one swimming in an axion cloud.
• Pinpointing the Particles: If they find this signal, they can work backward to figure out the exact mass of the axion and how strongly it interacts with itself. They estimate they could measure these properties with an accuracy of a few percent.

6. The Big Picture

The paper concludes that we don't need to find axions by smashing particles in a lab or looking for them in stars. Instead, we can find them by listening to the "music" of black holes colliding.

If LISA hears a black hole binary spiraling in a way that suggests it's dragging through a thick, invisible fog, it could be the first direct proof that these mysterious "axion-like" particles exist and that they have the specific self-interactions described in this model. It turns the universe's most violent events into a laboratory for testing the smallest particles.

Technical Summary

Technical Summary: Impact of Axion-Like Self-Interactions in Gravitational Atoms for LISA

Problem Statement
The microscopic nature of Dark Matter (DM) remains an open question, with ultralight bosonic particles (masses $10^{-24}$ to $1$ eV) representing a compelling candidate class. While previous studies have examined DM overdensities (spikes or minispikes) around black holes (BHs) using phenomenological profiles or superradiant instability mechanisms, these often rely on static assumptions or specific microphysical priors. Superradiant clouds require rotating Kerr BHs and typically populate excited states ($n=1, l=1$), whereas alternative formation mechanisms driven by self-interactions may produce different macroscopic states. This work addresses the need to understand the dynamical formation of "gravitational atoms" (GAs)—bosonic condensates around BHs—driven specifically by axion-like particle (ALP) self-interactions, and quantifies their observable imprint on Extreme and Intermediate Mass Ratio Inspirals (EMRIs/IMRIs) detectable by the Laser Interferometer Space Antenna (LISA).

Methodology
The authors adopt a dynamical formation mechanism for GAs proposed in [20], where the halo grows from an ambient DM distribution via a self-interaction-driven capture process, rather than assuming a pre-existing static overdensity.

• Theoretical Framework: The study models a scalar field with a periodic potential (Eq. 1) expanded to include quartic self-interactions (Eq. 2). In the Newtonian regime, this reduces to the Gross-Pitaevskii equation (Eq. 9) with a gravitational potential dominated by the central BH.
• Formation Dynamics: The growth of the GA is treated perturbatively. The rate of change in the halo mass depends on direct capture and stimulated capture (Bose enhancement) versus stripping. The system evolves toward a critical density $\rho_{crit}$ where self-interaction energy becomes negligible compared to gravity, leading to a saturated cloud (Eq. 21).
• Binary Evolution: The authors model an EMRI/IMRI where a secondary BH ($m$) inspirals through the GA overdensity surrounding a primary BH ($M_{BH}$). The orbital evolution is governed by an energy balance equation (Eq. 25) accounting for:
  1. Gravitational Wave (GW) emission (quadrupole formula).
  2. Dynamical Friction (DF) caused by the wake of the DM cloud.
  3. Accretion (ACC) of DM onto the secondary.
     The analysis shows DF dominates over ACC (Eq. 32), allowing the latter to be neglected in the phase evolution.
• Numerical Implementation: The GW frequency evolution is solved numerically (Eq. 33) using an adaptive step-size integrator. The phase accumulation is computed to determine the dephasing ($\delta\phi$) relative to a vacuum inspiral. Detectability is assessed via the mismatch metric $M$ and the required Signal-to-Noise Ratio (SNR) for distinguishability (Eq. 43). Bayesian parameter estimation is performed using a Fisher-matrix-inspired likelihood to constrain ALP parameters ($m_{dm}, f_a$).

Key Contributions and Results

• Dephasing Signatures: The presence of a GA induces a characteristic dephasing in the GW waveform. This effect enters at a negative Post-Newtonian (PN) order of $-5.5$ PN (Eq. 63), distinct from standard vacuum terms, preventing it from being absorbed into a redefinition of the chirp mass.
• Parameter Space Constraints: For conservative background DM densities ($\rho_{dm} \sim 10^3\text{--}10^4 \text{ GeV/cm}^3$) and primary BH masses $M_{BH} \sim 10^4\text{--}10^5 M_\odot$, LISA can probe:
  • ALP masses: $m_{dm} \sim 10^{-17}\text{--}10^{-15}$ eV.
  • Decay constants: $f_a \sim 10^{10}\text{--}3.2 \times 10^{12}$ GeV.
  • These constraints assume the GA reaches sufficient densities many cycles before merger.
• Dependence on Formation Timescales: The detectability is highly sensitive to the formation timescale $\tau_{crit}$. The parameter space exhibits a triangular morphology (Figs. 6, 7) where dephasing is maximized only when the formation time allows the cloud to reach high densities within the observation window.
• Parameter Recovery: For a binary with $M_{BH} \sim 10^4 M_\odot$, $\rho_{dm} = 10^3 \text{ GeV/cm}^3$, and SNR $\sim 20$, the authors find that specific ALP parameters ($m_{dm} \sim 2.5 \times 10^{-16}$ eV, $f_a \sim 6.3 \times 10^{10}$ GeV) maximize dephasing. In these high-mismatch regions, parameters can be recovered at the percent level (e.g., $\sim 10\%$ uncertainty for SNR=20, improving to $\sim 2\%$ for SNR=70).
• Scalability: Allowing for higher background densities ($\rho_{dm} \sim 10^5\text{--}10^7 \text{ GeV/cm}^3$) and higher mass BHs extends the accessible parameter space to $m_{dm} \sim 4.1 \times 10^{-19}$ eV and $f_a \sim 9.77 \times 10^{14}$ GeV.

Significance and Claims
The paper demonstrates that LISA can place constraints on axion-like particle masses and self-interactions without requiring additional couplings to Standard Model fields (e.g., photon couplings). The study highlights that the dynamical formation mechanism produces GAs in the fundamental state ($n=1, l=0$) with density profiles distinct from superradiant clouds ($n=2, l=1$) or static spikes.

The authors assert that while the maximum cloud masses in their dynamical model are small ($\sim 10^{-6} M_{BH}$), justifying the neglect of the cloud's self-gravity, the resulting environmental effects are sufficient to produce observable dephasing in the LISA band. The work establishes that LISA is uniquely suited to distinguish these waveforms from vacuum inspirals for SNRs $\lesssim 100$, providing a direct probe of the microphysical properties of ultralight bosons through their gravitational influence alone. The results are presented as consistent with, but physically distinct from, previous equilibrium cloud models, offering a more realistic, time-dependent scenario for testing DM candidates.