Probing small-scale anisotropic inflation with stochastic gravitational-wave background

This paper investigates how anisotropic primordial power spectra from specific inflation models, such as Finslerian and gauge field inflation, influence scalar-induced gravitational waves, concluding that current pulsar timing array observations cannot yet rule out the existence of small-scale anisotropic primordial perturbations.

The Gist

The Big Picture: Listening to the Universe's Hum

Imagine the universe is a giant, cosmic concert hall. For a long time, we thought the music playing there was perfectly uniform—a smooth, steady hum coming from all directions equally.

In June 2023, a group of astronomers using Pulsar Timing Arrays (PTA) (which are like ultra-precise cosmic metronomes made of spinning stars) announced they finally heard a low-frequency "hum" in the background. This is called the Stochastic Gravitational Wave Background (SGWB). It's the sound of the universe vibrating.

The big question is: What is making this noise?

• Option A: Supermassive black holes dancing in pairs (like two whales singing to each other).
• Option B: Ripples from the very beginning of time, created when the universe was a tiny, hot speck (called Scalar-Induced Gravitational Waves or SIGWs).

This paper focuses on Option B. The authors ask: What if the "music" from the beginning of the universe wasn't perfectly smooth? What if it had a slight "tilt" or directionality?

---

The Core Concept: The "Directional" Ripple

1. The Isotropic Assumption (The Perfect Sphere)

Usually, scientists assume that the energy from the Big Bang was distributed like a perfect, smooth sphere. No matter which way you look, the "loudness" is the same. This is called isotropic.

2. The Anisotropic Reality (The Tilted Sphere)

The authors propose a different idea: Anisotropy. Imagine the universe's early energy wasn't a perfect sphere, but a slightly squashed or tilted balloon.

• Analogy: Think of a drum. If you hit it dead center, the sound waves travel out evenly in a circle (isotropic). But if you hit it near the edge, or if the drum skin is stretched tighter on one side, the sound waves travel differently depending on the direction (anisotropic).

The paper investigates what happens if the "drum skin" of the early universe was stretched unevenly. This would create gravitational waves that are louder in some directions than others.

---

The Challenge: The "Foggy Glasses" Problem

Here is the tricky part: We can't see the directionality.

The gravitational waves the PTA detects come from incredibly small scales in the early universe. But our current detectors are like someone wearing thick, foggy glasses. They can hear the volume of the hum, but they can't tell if the sound is coming from the left or the right. The signal is an average of everything mixed together.

The Paper's Solution:
Even though we can't see the direction, the direction still changes the volume.

• Analogy: Imagine you are in a room with a speaker playing music. If the speaker is pointing directly at you, it sounds loud. If it's pointing away, it sounds quiet. Even if you can't see the speaker (because of the fog), if you measure the volume, you can guess where it's pointing.

The authors did the math to show that if the early universe had this "tilt" (anisotropy), the total volume of the gravitational wave hum we hear today would be different than if the universe were perfectly smooth.

---

The Investigation: Testing the Models

The authors tested two specific theories about how this "tilt" could have happened:

1. The "Gauge Field" Model (The Magnetic Tilt): Imagine invisible magnetic fields during the Big Bang that pulled the universe in one direction, stretching the fabric of space unevenly.
2. The "Finslerian" Model (The Lopsided Geometry): Imagine the rules of geometry themselves were slightly broken or "tilted" in the early universe, making space behave differently depending on which way you looked.

They ran simulations to see: If the universe was tilted like this, would the volume of the gravitational waves match what the PTA is hearing right now?

---

The Findings: What Did They Discover?

1. We Can't Rule It Out Yet: The current data from the PTA is not precise enough to say, "Yes, the universe is tilted" or "No, it's perfectly round." The "foggy glasses" are too thick. The data allows for both a perfectly smooth universe and a slightly tilted one.
2. The "Volume" Clue: However, the tilt does change the predicted volume of the waves. The authors calculated exactly how much the volume would change for different amounts of tilt.
3. The Future is Bright (LISA): The paper suggests that while current PTA data is too blurry to see the tilt, future detectors like LISA (a space-based gravitational wave observatory) might be able to.
   • Analogy: It's like upgrading from a standard-definition TV to 4K. The PTA is the old TV; it hears the noise but can't see the details. LISA will be the 4K TV, potentially sharp enough to see if the "drum" was tilted.

The Conclusion in Plain English

The universe might have a "preferred direction" from its birth, making the gravitational waves it emits slightly louder in some directions than others.

• Current Status: Our current detectors (PTA) are too blurry to confirm this. They can't tell the difference between a perfectly round universe and a slightly squashed one.
• The Good News: The math shows that this "squash" leaves a fingerprint on the loudness of the waves.
• The Future: By combining current data with future observations (like LISA) and looking at other cosmic clues (like the Cosmic Microwave Background), we might eventually be able to prove that the early universe wasn't perfectly symmetrical, but had a subtle, directional tilt.

In short: The universe might be a bit "lopsided" at its core, and this paper provides the mathematical map to find out if we're right, once our cosmic ears get sharp enough to hear the difference.

Technical Summary

1. Problem Statement

In June 2023, multiple Pulsar Timing Array (PTA) collaborations (NANOGrav, EPTA, PPTA, CPTA) reported evidence for a stochastic gravitational-wave background (SGWB) in the nanohertz frequency range. While the astrophysical origin (supermassive black hole binaries) is a leading candidate, cosmological origins, particularly Scalar-Induced Gravitational Waves (SIGWs) generated by large-amplitude primordial curvature perturbations on small scales, are also strongly favored by Bayesian analysis.

Current theoretical models of SIGWs typically assume a statistically isotropic primordial power spectrum (PPS), meaning the spectrum depends only on the magnitude of the wavenumber $k$, not its direction. However, many inflationary models (e.g., those involving vector/gauge fields or Finslerian spacetime backgrounds) predict anisotropic primordial perturbations.

• The Gap: Current PTA data has limited angular resolution and cannot directly resolve anisotropies in the SGWB signal on small scales.
• The Question: Can current PTA observations constrain the existence of small-scale anisotropic primordial perturbations, and how do these anisotropies affect the observable energy density spectrum of SIGWs?

2. Methodology

The authors employ a model-independent parametric approach combined with specific inflationary model analysis to investigate this problem.

A. Theoretical Framework

1. Second-Order SIGW Calculation: The authors review the generation of second-order SIGWs during the radiation-dominated (RD) era. They derive the energy density spectrum $\Omega_{GW}(k)$ from the primordial curvature perturbation power spectrum $P_\zeta(k)$.
2. Anisotropic Power Spectrum Parameterization: Instead of assuming isotropy, they model the PPS as:
   $$P_\zeta^{\hat{n}}(k) = P_{0,\zeta}(k) \sum_{l=0}^{4} C_l P_l(\hat{n} \cdot \hat{k})$$
   where $P_l$ are Legendre polynomials, $\hat{n}$ is the anisotropy direction, and $C_l$ are anisotropy parameters ($l=1, 2, 3, 4$).
3. Spatial Averaging: Since current detectors cannot resolve small-scale anisotropy, the authors perform a spatial average over the direction $\hat{n}$ to derive the effective isotropic energy density spectrum observed by PTAs. This averaging retains the influence of the anisotropy parameters $C_l$ on the total amplitude.
4. Multipole Expansion: They also derive the multipole moments ($\Omega_l$) of the SIGW energy density spectrum to characterize the residual anisotropy at small scales, which could be probed by future high-precision detectors.

B. Specific Models Considered

The study focuses on two specific classes of anisotropic inflation:

1. Gauge Field Inflation: Introduces a vector field during inflation, leading to a PPS with a quadrupole anisotropy ($C_2 \neq 0$).
2. Finslerian Inflation: Modifies the spacetime background to Finsler geometry, leading to a PPS with a dipole anisotropy ($C_1 \neq 0$).

C. Data Analysis and Constraints

1. PTA Data: They utilize the NANOGrav 15-year dataset free spectra (Kernel Density Estimator representations) to construct likelihood functions.
2. Bayesian Inference: Using the bilby package and dynesty nested sampler, they perform Bayesian analysis to constrain the parameters of the PPS ($A_\zeta, \sigma, f_*$) and the anisotropy parameters ($C_1, C_2$).
3. Multi-Source Constraints: They combine PTA data with large-scale cosmological constraints from CMB, BAO, and Big-Bang Nucleosynthesis (BBN). These constraints limit the total energy density of the SGWB to avoid altering the effective number of relativistic species ($N_{eff}$).
4. Future Projections: They calculate the Signal-to-Noise Ratio (SNR) for the LISA mission to assess the detectability of these models in the mHz frequency band.

3. Key Contributions

• Derivation of Anisotropic SIGW Spectra: The paper provides explicit analytical expressions for the energy density spectrum of second-order SIGWs generated by a general anisotropic primordial power spectrum (up to $l=4$) after spatial averaging.
• Model-Independent Constraints: It establishes a framework to constrain anisotropy parameters ($C_l$) without committing to a specific inflationary Lagrangian, using a phenomenological expansion.
• Bayesian Model Comparison: The authors perform a rigorous Bayesian comparison between isotropic SIGW models, anisotropic SIGW models, and hybrid models (SIGW + Supermassive Black Hole Binaries).
• Interplay of Frequency Bands: The study analyzes the complementarity between PTA (nHz) and LISA (mHz) observations in constraining the parameter space of anisotropic inflation.

4. Results

• PTA Constraints on Anisotropy: Current PTA data (NANOGrav 15-year) can effectively constrain the isotropic parameters of the PPS (amplitude $A_\zeta$, width $\sigma$, peak frequency $f_*$). However, it cannot effectively constrain the anisotropy parameters $C_1$ and $C_2$.
  • Reason: There is a strong parameter degeneracy between the isotropic amplitude and the anisotropy parameters in the energy density spectrum. The current data precision is insufficient to break this degeneracy.
• Bayes Factors:
  • The Bayes factor for SIGW models (both isotropic and anisotropic) relative to the SMBHB model is high ($\sim 10$), suggesting SIGWs are the likely source of the PTA signal.
  • The Bayes factor comparing the anisotropic model ($C_2 \neq 0$) to the isotropic model ($C_1=C_2=0$) is approximately 1.08. This indicates that current data cannot distinguish between isotropic and anisotropic primordial power spectra; the data is consistent with both.
• Large-Scale Constraints: Combining PTA data with CMB/BAO/$N_{eff}$ limits provides tighter bounds on the parameter space ($A_\zeta$ vs. $C_l$), but the anisotropy parameters remain largely unconstrained due to the degeneracy.
• LISA Projections:
  • If SIGWs dominate the PTA signal, the peak frequency $f_*$ is constrained to $\sim 10^{-6.7}$ Hz.
  • In this scenario, the resulting SIGW signal in the LISA frequency band is too weak ($SNR < 10^{-2}$) to be detected, regardless of the values of $C_1$ and $C_2$.
  • Conversely, if the peak frequency were high enough to be detected by LISA ($f_* > 10^{-3}$ Hz), the signal would likely be too strong for PTA observations.
  • Conclusion: The specific anisotropic models considered cannot simultaneously dominate PTA observations and produce a detectable signal in LISA.
• Specific Model Limits: The authors note that specific theoretical models (e.g., the Yokoyama-Soda model) often impose intrinsic theoretical bounds on the anisotropy parameters (e.g., $g_* \ll 1$), which are stricter than current observational limits.

5. Significance

• Validation of Isotropy Assumption: The study confirms that the standard assumption of isotropy in PPS analysis is statistically valid given current data, as anisotropic models are not ruled out but also not preferred over isotropic ones.
• Future Prospects: It highlights that current cosmological observations are insufficient to probe small-scale anisotropic inflation. Future constraints will require:
  1. Higher precision PTA data to break parameter degeneracies.
  2. Multi-band GW observations (combining PTA, LISA, and potentially future detectors like Einstein Telescope) to cover a wider range of frequencies.
  3. Theoretical priors from specific inflationary models to restrict the parameter space of $C_l$.
• Theoretical Guidance: The derived analytical expressions for anisotropic SIGWs provide a necessary toolkit for future studies aiming to detect or constrain primordial anisotropy as detector sensitivity improves.

In summary, while the paper establishes that small-scale anisotropic inflation is a viable explanation for the PTA signal, it concludes that current data cannot distinguish it from isotropic inflation, and future multi-messenger GW astronomy is required to probe these anisotropic features.