On-Shell Bootstrap of Loop Inflation Correlators with Spectral Dispersion

This paper introduces a "spectral dispersion" bootstrap strategy that combines de Sitter spectral decomposition with dispersion relations to efficiently compute loop-level cosmological correlators by reconstructing them from on-shell nonlocal signals and quasinormal modes.

The Gist

Imagine the universe as a giant, expanding drum. When it was very young, during a period called "inflation," it expanded so fast that tiny quantum ripples were stretched out into huge waves. These waves left behind a faint pattern in the cosmic microwave background, like the grooves on a vinyl record. Scientists want to read these grooves to learn about heavy particles that existed back then, particles that are too heavy to be made in any particle accelerator on Earth.

This paper introduces a new, clever way to "read" these cosmic grooves, specifically looking at complex patterns created by loops of particles. The authors call their method "Spectral Dispersion."

Here is a simple breakdown of how it works, using everyday analogies:

1. The Problem: The Cosmic "Black Box"

Usually, to understand what happened inside a complex machine, you have to take it apart and look at every tiny gear. In physics, calculating how these heavy particles interact involves incredibly difficult math with many layers of time and space. It's like trying to predict the exact sound of a symphony by calculating the vibration of every single molecule in every instrument simultaneously. It's possible, but it's a nightmare.

2. The Insight: Listening to the "Echoes"

The authors realized they don't need to calculate every single gear. Instead, they can listen to the echoes.

In the expanding universe, when heavy particles pop into existence and then disappear, they leave behind a specific "signature" or "echo" in the cosmic data. The authors call this the "nonlocal signal."

• The Analogy: Imagine you are in a large canyon. You clap your hands (the interaction). You hear the direct sound, but you also hear the echo bouncing off the walls. The echo tells you about the shape of the canyon and the distance to the walls, without you needing to measure the walls directly.
• In this paper, the "echo" is the part of the data that comes from particles that briefly existed "on-shell" (meaning they behaved like real, physical particles for a moment before vanishing).

3. The Method: Spectral Dispersion

The authors combine two powerful ideas to turn these echoes into a full picture:

• Spectral Decomposition (The Prism): Imagine shining white light through a prism. It splits into a rainbow of distinct colors (frequencies). Similarly, the authors realized that the complex "echo" of a loop of particles isn't just one messy sound; it is actually a sum of many distinct, pure tones (called "quasinormal modes"). Each tone corresponds to a specific way the particle can vibrate or decay.
• Dispersion Relations (The Reconstruction): In physics, if you know the "echoes" (the non-analytic parts) of a signal, you can mathematically reconstruct the entire signal, provided you know the rules of the game (analyticity). It's like knowing the specific frequencies of a song allows you to write down the entire sheet music, even the parts you didn't hear directly.

The "Spectral Dispersion" Strategy:

1. Identify the Echoes: Calculate the "nonlocal signal" (the echo) for the simplest version of the interaction.
2. Split the Echo: Use the "prism" (spectral decomposition) to break that echo down into a list of pure tones (modes).
3. Reconstruct the Whole: Use the "reconstruction rule" (dispersion) to turn those pure tones back into the full, complex result.

4. What They Did

The authors used this method to solve problems that were previously very hard to calculate. They looked at specific scenarios where particles form a "bubble" loop (a particle goes around in a circle before disappearing).

• They calculated these loops for scalar particles (like simple dots) and vector particles (like arrows with direction).
• They handled cases where the particles interact directly and cases where they interact through movement (derivatives).
• The Result: They produced new, much simpler formulas for these complex cosmic patterns.

5. The "Glitch" (Renormalization)

There is one catch. When you reconstruct the song from the echoes, you might get a few extra notes that don't belong to the original melody. In physics, these are called "local counterterms."

• The Analogy: Imagine you are trying to reconstruct a song from an echo, but your microphone also picked up some static noise. You can hear the song perfectly, but you have to manually decide how to filter out the static.
• The authors show that their method gives you the "song" (the physical prediction) perfectly, but the "static" (the part that depends on how you set up your math) needs to be fixed by a standard rule called a "renormalization condition." Once you fix that, the rest of the result is a solid, unchangeable prediction.

Summary

This paper is like a new toolkit for cosmologists. Instead of trying to build a complex machine from scratch (doing the hard math from the beginning), they showed you how to listen to the machine's hum (the on-shell data), break that hum down into simple notes, and then use those notes to write down the entire blueprint of the machine. This makes it much faster and easier to predict what the universe should look like if heavy, exotic particles existed during inflation.

Technical Summary

Technical Summary: On-Shell Bootstrap of Loop Inflation Correlators with Spectral Dispersion

Problem Statement
The computation of cosmological correlators in inflationary spacetime, particularly those involving massive particle exchanges, is central to "Cosmological Collider" (CC) physics. While tree-level massive correlators have been successfully computed using techniques like differential equations and family-tree decompositions, loop-level calculations remain technically challenging due to multi-layered, nested time integrals over products of special functions. Existing results for loop processes are sparse, largely limited to specific 1-loop bubble graphs. The authors identify a need for a more efficient, analytic approach to compute massive inflation correlators at the loop level, which are crucial for probing heavy particles with masses around or beyond the Hubble scale ($H$).

Methodology: Spectral Dispersion
The paper proposes a new bootstrap strategy termed spectral dispersion, which synthesizes two established techniques: dispersion relations and spectral decomposition. The method relies on the specific analytic structure of inflation correlators in de Sitter (dS) space.

1. Analyticity and On-Shell Data:

   • Unlike flat-space scattering amplitudes where on-shell intermediate states generate poles, on-shell states in dS space generate branch points (nonlocal signals) due to the dilution of particles by Hubble expansion.
   • The authors utilize the observation that the full correlator is analytic everywhere except for these branch cuts. The discontinuity across these cuts is fully determined by "on-shell data," specifically the nonlocal signals arising from intermediate particles going on-shell.
   • A line dispersion relation is employed, expressing the full correlator as an integral over the discontinuity of a chosen branch cut (typically in the complex momentum plane), up to local contact terms.

2. Spectral Decomposition:

   • The authors apply the dS spectral decomposition (the dS counterpart of the Källén-Lehmann representation) to 1-loop bubble graphs. This technique rewrites a loop correlator as a spectral integral over the mass of a massive propagator, weighted by a spectral function $\rho(\nu)$.
   • Crucially, for bubble graphs, the spectral integral over mass translates into a discrete sum over "quasinormal modes" (states with definite scaling dimensions $\Delta$). The nonlocal signal of a 1-loop bubble is shown to be a discrete sum of nonlocal signals of tree-level graphs exchanging these quasinormal modes.

3. The Bootstrap Strategy:

   • The core conjecture (verified in the paper) is that the full 1-loop correlator $J$ can be reconstructed from its nonlocal signal $J_{NS}$ by applying the line dispersion relation to each term in the discrete spectral sum.
   • Specifically, if the nonlocal signal is $J_{NS} = \sum C_{n,\Delta} I^{\Delta+2n}_{NS}$ (where $I$ is the tree-level correlator for a mode with dimension $\Delta$), the full correlator is conjectured to be $J = \sum C_{n,\Delta} I^{\Delta+2n} + \text{local terms}$.
   • The "local terms" represent UV divergences and contact interactions that cannot be fixed by analyticity alone and must be determined by renormalization conditions.

Key Contributions and Results
The authors apply this spectral dispersion method to bootstrap 1-loop bubble correlators for both 3-point and 4-point functions involving massive scalars and vector bosons.

• Scalar and Vector Bubbles: The method is applied to:
  • Directly coupled massive scalar bubbles ($\sigma^2 \phi_c^2$).
  • Derivatively coupled massive scalar bubbles ($(\nabla \sigma)^2 \phi_c^2$).
  • Massive vector boson bubbles ($A^2 \phi_c^2$).
• Explicit Results: The paper provides explicit, simplified analytic expressions for these correlators. The results are presented as a sum of:
  • Nonlocal Signals: Oscillatory features ($k^{\Delta}$) arising from on-shell production.
  • Local Signals: Non-oscillatory features associated with specific kinematic limits.
  • Background: Regular terms, split into renormalization-dependent pieces (UV divergent sums) and renormalization-independent "irreducible" backgrounds.
• Handling Divergences: A significant technical achievement is the systematic identification of UV divergences within the spectral sum. The authors demonstrate that divergent terms in the $n$-summation (from the spectral decomposition) correspond exactly to the kinematic structures of local counterterms required for renormalization. This allows them to isolate the finite, renormalization-independent predictions of the bootstrap.
• Applications to Primordial Spectra: The results are applied to compute the 1-loop primordial bispectrum and trispectrum for inflaton fluctuations. The authors provide limiting expressions for the squeezed limit ($k_{small} \ll k_{large}$) and the large-mass limit, showing how the oscillatory signatures of heavy particles manifest in these spectra.

Significance and Claims
The paper claims that the spectral dispersion technique offers a "simple and intuitive" alternative to direct diagrammatic integration for loop correlators. Its significance lies in:

1. Simplification: It yields results in a much simpler form compared to previous purely spectral methods, explicitly revealing the spectral structure (discrete sum over scaling dimensions) of loop processes.
2. Generality: The method is applicable to arbitrary spins and couplings, provided the loop lines satisfy dS-covariant dispersion relations. It handles both covariant and non-covariant (boost-breaking) interactions by utilizing tensor embedding formalisms.
3. On-Shell Philosophy: It extends the on-shell bootstrap philosophy (successful in flat-space amplitudes) to dS loop processes. By relying solely on on-shell data (nonlocal signals) and analyticity, it separates physical propagation effects from local EFT counterterms cleanly.
4. New Results: The paper presents new analytical results for derivatively coupled scalar loops and massive vector loops, which were previously difficult to compute.

The authors maintain a modest stance, noting that while the method simplifies existing calculations and provides new results for specific topologies (bubbles), it represents a "first use" of the technique. They acknowledge that extending this to more complex topologies (triangles, boxes) or breaking dS isometries more severely remains a challenge for future work.