⚛️ general relativity

Split Representations and Bubble Resummation for Massive de Sitter Correlators

This paper introduces a new mathematical framework combining spectral and split representations to factorize multi-loop diagrams in de Sitter space, enabling the resummation of massive scalar loop contributions and the direct identification of cosmological collider signals.

Original authors: Jonathan Gräfe, Ivo Sachs

Published 2026-02-11

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Jonathan Gräfe, Ivo Sachs

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

Imagine you are trying to listen to a very faint, specific melody being played by a single violin in the middle of a massive, roaring thunderstorm.

In this analogy, the "thunderstorm" is the expanding universe (specifically, a mathematical model called de Sitter space). The "violin melody" represents the tiny quantum signals left over from the very beginning of time—signals that physicists hope to study to understand how the universe was born.

This paper, written by Jonathan Gräfe and his colleagues, is essentially a new, high-tech "noise-canceling headphone" and "signal booster" for physicists. Here is the breakdown of how it works.

1. The Problem: The Cosmic Static

When physicists try to calculate how particles interacted in the early universe, they run into a mathematical nightmare. In a flat, still space (like a quiet room), math is easy because everything stays put. But in an expanding universe, space itself is stretching and pulling everything apart.

This stretching makes the math "tangled." If you try to calculate a complex interaction (a "multi-loop diagram"), the equations become so messy and nested that they are almost impossible to solve. It’s like trying to untangle a massive ball of yarn where every thread is also moving and stretching.

2. The Solution: The "Split" Strategy

The authors introduce a clever trick called the Split Representation.

Imagine you have a giant, complicated knot of yarn. Instead of trying to untangle the whole knot at once, the "Split" method allows you to pretend the knot is actually made of many smaller, simpler loops that are just temporarily stuck together.

By "splitting" the complex diagrams into simpler "building blocks," they can solve the easy parts first and then use a mathematical "glue" to put them back together. This turns a problem that was once a mountain into a series of small, manageable hills.

3. The "Bubble Chain" Resummation

The paper specifically looks at something called "Bubble Chains."

Imagine a chain of soap bubbles. In the quantum world, particles can pop in and out of existence, creating "bubbles" of energy. In a high-energy environment, you don't just get one bubble; you get a massive, infinite chain of them.

Previously, physicists could only calculate one or two bubbles at a time. If they tried to add more, the math exploded. The authors developed a way to "resum" these bubbles. This is like finding a magic formula that tells you the total weight of an infinite chain of bubbles without having to weigh every single one individually. They found a way to sum up the entire infinite chain into one clean, elegant equation.

4. Why does this matter? (The Cosmological Collider)

Why go to all this trouble? Because of something called the Cosmological Collider.

Even though we can't build a particle accelerator as big as the universe here on Earth, the early universe was a massive particle accelerator. The "violin melody" (the signal) carries information about heavy, mysterious particles that existed billions of years ago.

By using these new mathematical tools, the authors can:

Filter the noise: They can separate the "background roar" of the universe from the "musical signal" of the particles.
Predict the tune: They can tell exactly what "frequency" (mass and spin) a particle would have based on the patterns in the cosmic background.
See the "Flow": They can track how these signals change as the universe evolves, much like watching how a ripple moves across a pond.

Summary

In short: The universe is a noisy, stretching, expanding mess. This paper provides a new mathematical toolkit to split complex problems into simple pieces and sum up infinite chains of quantum events. This allows us to peer through the cosmic static to hear the "music" of the particles that built our universe.

Technical Summary: Split Representations and Bubble Resummation for Massive de Sitter Correlators

Authors: Jonathan Gräfe and Ivo Sachs
Field: Theoretical High-Energy Physics / Cosmology (Quantum Field Theory in Curved Spacetime)

1. The Problem: Complexity of Loop Calculations in de Sitter Space

The study of cosmological correlation functions is essential for "cosmological collider physics," which aims to use primordial fluctuations to probe high-energy particle physics. However, calculating these correlators at the multi-loop level in de Sitter (dS) space is notoriously difficult due to:

Lack of Time-Translation Symmetry: Unlike Minkowski space, energy is not conserved in an expanding universe, preventing the use of standard momentum-space factorization.
Complex Mode Functions: Propagators in dS involve Hankel functions, leading to nested time integrals that are mathematically cumbersome.
Non-Perturbative Effects: Standard perturbation theory fails to capture essential physics like vacuum stability and infrared singularities, which often require the resummation of infinite series of diagrams (e.g., bubble chains in large- $N$ models).

While progress has been made for conformally coupled or massless scalars, a general framework for massive bulk fields at loop order has been missing.

2. Methodology: The Split and Spectral Representation

The authors propose a novel method to factorize multi-loop momentum space diagrams by combining two mathematical frameworks:

The Split Representation: Based on harmonic analysis, this method decomposes bulk-to-bulk propagators into a spectral integral of products of bulk-to-boundary propagators (harmonic functions). This effectively "splits" a vertex into disconnected pieces, allowing for a more manageable diagrammatic structure.
Källén-Lehmann Spectral Representation: The authors use this to express loop integrals as spectral integrals over simpler tree-level diagrams.

Key Innovation: By applying the split representation to the Schwinger-Keldysh (in-in) formalism, the authors demonstrate that even though energy is not conserved, the spectral parameter ( $\tilde{\nu}$ ) in harmonic analysis acts as a conserved quantity. This allows for the factorization of $N$ -loop bubble chains into a geometric series, mirroring the simplicity of flat-space resummation.

3. Key Contributions and Results

A. Factorization and Resummation of Bubble Chains
The paper provides a rigorous derivation showing that a chain of $N$ bubble loops can be expressed as a single spectral integral where the spectral density is simply the $N$ -th power of the one-loop spectral density ( $\Sigma_\nu^N$ ). This allows the authors to perform a non-perturbative resummation of the entire bubble chain using a geometric series:
$I_\nu^{np} \propto \int d\tilde{\nu}' \frac{\text{Tree-level terms}}{1 + \lambda \Sigma_\nu(\tilde{\nu}')}$

B. Renormalization in dS
The authors identify the UV divergences within the spectral function itself. They implement a renormalization scheme (a modified minimal subtraction) by subtracting the divergent asymptotic part of the spectral series ( $P_\nu^{asy}$ ). This renders the spectral function finite and allows for a consistent treatment of higher-loop diagrams via local counterterms.

C. Analytical Results in $d=2$ and $d=3$

In $d=2$ : The authors find closed-form expressions for the spectral function in terms of digamma functions ( $\psi$ ). They describe the "pole flow"—how the poles of the resummed propagator move in the complex plane as the coupling $\lambda$ changes.
In $d=3$ (Physical Case): While a closed-form expression for the spectral function is not found, they provide a renormalized version and explicitly compute the one-loop bubble diagram. They successfully separate the result into an EFT background (analytic) and cosmological collider signals (oscillatory).

D. Cosmological Collider Signals
The authors show that the "signal" (the part of the correlator containing information about particle masses) can be read directly from the spectral function's poles. They confirm that these signals manifest as oscillations in the momentum ratios, $\propto \cos[2\tilde{\nu} \log r + \theta]$ , which is the hallmark of cosmological collider physics.

4. Significance and Implications

Theoretical Advancement: This work provides the first systematic method to resum massive scalar loops in de Sitter space, bridging the gap between perturbative results and non-perturbative large- $N$ physics.
Predictive Power: The "split representation" allows researchers to identify the frequency and amplitude of cosmological signals directly from the integrand, bypassing the need for tedious multi-dimensional integration.
Stability and Unitarity: The paper explores the limits of the theory, showing that for negative couplings ( $\lambda < 0$ ), the resummation reveals unphysical behaviors (loss of spectral positivity), which corresponds to vacuum instability and a breakdown of unitarity.
Universality: The method links cosmological correlators to the properties of conformal field theories (CFT), where the poles of the spectral function correspond to the scaling dimensions of operators in the dual theory.