🔭 astrophysics

Self-interacting neutrinos in cosmological perturbation theory -- integrating the collision kernel

This paper derives an exact analytic expression for the collision kernel coefficients in the Boltzmann hierarchy of self-interacting neutrinos by transforming the integration kernel into angular derivatives of a Yukawa potential, enabling a compact, exact rational-plus- $\pi^2$ implementation for cosmological perturbation codes like CLASS.

Original authors: Jakob K. Mogensen, Steen Hannestad, Thomas Tram

Published 2026-06-03

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jakob K. Mogensen, Steen Hannestad, Thomas Tram

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

The Big Picture: Neutrinos as a Crowded Dance Floor

Imagine the early universe as a massive, crowded dance floor. The dancers are neutrinos, tiny ghost-like particles that usually zip through everything without touching anyone. In our standard understanding of physics, they are so shy that they barely interact with each other; they just drift through the crowd (a process called "free-streaming").

However, this paper explores a "what if" scenario: What if these neutrinos had a secret way to bump into each other?

The authors are studying a model where neutrinos interact with each other via a heavy, invisible "messenger" particle (a massive scalar). When two neutrinos bump, they bounce off one another. This interaction changes how the neutrinos move and how they affect the rest of the universe, specifically the Cosmic Microwave Background (CMB)—which is essentially the "afterglow" or the oldest light in the universe, like a photograph of the dance floor taken billions of years ago.

The Problem: The "Mathematical Traffic Jam"

To predict what happens on this dance floor, scientists use a set of complex equations called the Boltzmann hierarchy. Think of these equations as a traffic report for the neutrinos.

To make the traffic report accurate, you have to calculate a specific "collision term." This term describes exactly how often and how hard the neutrinos bump into each other.

The Old Way: Previously, scientists had to approximate this collision term. They used a "Relaxation Time Approximation" (RTA). Imagine trying to predict traffic by guessing the average speed of cars rather than calculating every single car's path. It's a good guess, but it's not perfect. To get the numbers for this guess, they had to solve a triple integral (a math problem with three layers of complexity) that was notoriously unstable and prone to errors. It was like trying to balance a house of cards on a shaking table.
The Goal: The authors wanted to stop guessing. They wanted to solve the math problem exactly.

The Solution: A New Mathematical Shortcut

The authors, Jakob, Steen, and Thomas, found a way to solve this "traffic jam" of math perfectly. Here is how they did it, using an analogy:

The Messy Kernel: The collision term is like a very complicated, tangled knot of string. It has a lot of "noise" (high inverse powers of a variable $P$ ) that makes it hard to untangle.
The Magic Trick (Derivatives): The authors realized that this messy knot could be described as the result of taking derivatives (a calculus operation that measures how fast something changes) of a much simpler, smoother string (the Yukawa potential, $e^{-P/2}/P$ ).
Moving the Derivatives: Instead of trying to untangle the knot directly, they used a mathematical trick called "integration by parts." This allowed them to move the "derivative" operation from the messy knot onto the smooth string.
The Result: Once the derivatives were moved, the messy knot simplified into a few standard, manageable pieces. They reduced the entire complex problem into a single family of base integrals that follow a simple, repeating pattern (a recurrence relation).

The "Recipe" for Exact Answers

Because they found this pattern, the authors created a "recipe" (a set of exact formulas) to calculate the interaction strength for any level of detail (called multipoles, denoted by $\ell$ ).

The Output: They didn't just get a messy decimal number. They found that the answer is always a combination of rational numbers (clean fractions) and $\pi^2$ (a specific mathematical constant).
The Tool: They wrote a computer program (a Jupyter notebook) that uses "exact rational arithmetic." This means the computer doesn't round off numbers like a calculator does; it keeps them as perfect fractions until the very last step. This eliminates the "noise" and errors that plagued previous methods.

What Did They Find? (The Results)

The authors compared their new, exact numbers with the old, approximate numbers used in major cosmology software (like class and CAMB).

The Surprise: The difference between the old "guess" and the new "exact" answer is tiny.
The Observation: When they plugged these new numbers into simulations of the universe, the resulting "photograph" of the early universe (the CMB power spectrum) looked almost identical to the old one. The differences were so small (less than one part in a thousand) that they are currently unobservable with our telescopes.
The Conclusion: While the old method was an approximation, it was a very good one. However, the authors have now provided the exact mathematical foundation. It's like upgrading from a high-quality map to a perfect GPS coordinate. The route looks the same, but the underlying data is now mathematically flawless.

Summary in One Sentence

This paper provides a perfect, error-free mathematical recipe for calculating how neutrinos bump into each other in the early universe, proving that while our previous approximations were surprisingly accurate, we can now do the math exactly without any guesswork.

Technical Summary: Self-interacting neutrinos in cosmological perturbation theory - integrating the collision kernel

Problem Statement
Cosmological constraints on self-interacting neutrinos, mediated by a light scalar or pseudo-scalar, require solving the Boltzmann hierarchy with a collision term projected onto momentum-averaged multipoles. Previous treatments, particularly those utilizing the Relaxation Time Approximation (RTA) as in Ref. [23], replaced the complex collision term with effective momentum-averaged quantities (coefficients $\alpha_\ell$ ). However, the computation of these numerical front factors previously relied on solving a highly unstable triple integral. While the RTA was found to be valid to excellent precision, the numerical instability of the integral calculation posed a challenge for obtaining exact analytic expressions for the coefficients $\alpha_\ell$ .

Methodology
The authors revisit the collision kernel for neutrino-neutrino scattering ( $\nu\nu \to \nu\nu$ ) mediated by a very massive scalar $\phi$ . In this limit, the mediator population is Boltzmann-suppressed, and the interaction is confined to the neutrino sector. The paper derives an exact analytic expression for the multipole integral determining the coefficients $\alpha_\ell$ .

The core methodology involves a three-step analytic reduction of the collision kernel:

Kernel Decomposition: The complex integration kernel is expressed as angular derivatives of a base kernel, specifically the Yukawa-potential form $e^{-P/2}/P$ , where $P$ is the momentum transfer magnitude.
Integration by Parts: The angular derivatives are moved onto the Legendre polynomials via integration by parts. This reduces the angular dependence to three finite angular mixing operators ( $O^{(1)}_\ell, O^{(2)}_\ell, O^{(3)}_\ell$ ) and generates specific boundary terms.
Recurrence Reduction: The remaining momentum integrals are reduced to a single base family, $I_{4,4}^k$ , which obeys a stable first-order recurrence relation. The base integral $I_{4,4}^k$ is shown to have a closed form involving rational numbers and $\pi^2$ .

The authors utilize exact rational arithmetic to compute the coefficients, representing the final result as a sum of a rational part and a $\pi^2$ part ( $R_\ell + S_\ell\pi^2$ ). This avoids the numerical instabilities associated with direct integration.

Key Contributions

Exact Analytic Expression: The paper provides an exact analytic formula for the collision kernel integral $I_\ell$ (Eq. 4.27), replacing the previous unstable numerical integration.
Rational-Plus- $\pi^2$ Representation: It is demonstrated that for all multipoles $\ell$ , the integral $I_\ell$ (and consequently $\alpha_\ell$ ) can be expressed exactly as the difference between an integer (or rational number) and another integer (or rational) multiplied by $\pi^2$ .
Recurrence Relations: The authors derive stable recurrence relations for the base momentum integrals and the angular weights ( $W^{(i)}_{\ell k}$ ), allowing for efficient computation of coefficients for arbitrarily high multipoles.
Implementation: A public Jupyter notebook (IntegralComputation.ipynb) is provided, implementing the exact rational arithmetic and a Chudnovsky series approximation for $\pi^2$ to convert results to double precision. An asymptotically constrained rational approximation suitable for Boltzmann codes like CLASS is also provided.

Results

Coefficients $\alpha_\ell$ : The authors calculate exact values for $\alpha_\ell$ up to $\ell=10$ (Table 1). They confirm that $\alpha_0$ and $\alpha_1$ vanish, consistent with energy and momentum conservation.
Asymptotic Behavior: The coefficients approach an asymptotic value of $\alpha_\ell \to 35\pi/216 \approx 0.50905$ for $\ell \gtrsim 26$ .
Numerical Impact: When implemented in the CLASS code, the exact $\alpha_\ell$ values are compared against the approximate values from Ref. [23]. The results show that the differences in the angular temperature power spectrum ( $C_\ell^{TT}$ ) are at the level of one per mille ( $10^{-3}$ ) or less.
Observables: Plots of density ( $\delta_\nu$ ) and velocity ( $\theta_\nu$ ) transfer functions show no visually distinct differences between the exact and approximate prescriptions.

Significance and Claims
The paper claims that while the analytical integration provides an exact solution to the previously numerically unstable problem, the practical impact on cosmological observables is modest. The authors state that the corrections to the $\alpha_\ell$ coefficients lead to "minute changes" in neutrino observables such as $\delta_\nu$ , $\theta_\nu$ , and $\sigma_\nu$ .

The significance of the work lies primarily in the mathematical rigor and stability of the solution:

It replaces an unstable numerical procedure with a stable, exact analytic framework based on rational arithmetic.
It provides a definitive benchmark for the coefficients used in cosmological Boltzmann codes.
The authors note that while the changes are likely unobservable given current data precision, the exact treatment removes theoretical uncertainty regarding the numerical factors. They also suggest that similar analytical treatments could be explored for other processes, such as inverse neutrino decays, where the coefficients are significantly larger.

The paper concludes that the exact integration is a technical improvement that validates the existing RTA approach used in cosmological analyses, rather than a discovery of new physical phenomena or a requirement for immediate revision of cosmological constraints.