⚛️ general relativity

Entropy production during reheating after inflation and General Relativistic Entropic Acceleration

This paper quantifies the irreversible entropy production during cosmological preheating using lattice simulations and demonstrates that while this entropy serves as a robust diagnostic for dynamical regimes, its feedback into the expansion rate via General Relativistic Entropic Acceleration is substantial only when coherent energy efficiently fragments into radiation, rather than simply when entropy is copiously produced.

Original authors: Juan Garcia-Bellido

Published 2026-09-18

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Juan Garcia-Bellido

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 universe began in a state of unimaginable density and heat, but before it could become the warm, star-filled cosmos we see today, it had to pass through a violent, chaotic transition. This period, known as reheating, is the bridge between the rapid, smooth expansion of the early universe and the hot, particle-filled era that followed. Imagine the universe as a giant, silent drum that has just been struck; the energy stored in that initial strike must be transferred into the vibration of countless tiny particles to create the matter and radiation that fill space. For decades, physicists have understood that this transfer involves a process called preheating, where the energy of the early universe's dominant field explodes into a sea of new particles. However, while scientists have long tracked the energy and the speed of these particles, they have paid less attention to the disorder, or entropy, created during this explosion. Entropy is a measure of how spread out energy is; a single, organized wave has low entropy, while a chaotic jumble of many waves has high entropy. Understanding exactly how much disorder is created, and how fast, is crucial because this process dictates how the universe expands and cools in its earliest moments.

A researcher has now taken a direct look at this hidden disorder by running massive computer simulations of the early universe. They focused on three different scenarios for how the initial energy might have been transferred to new particles, using a powerful software tool designed to model the behavior of fields in an expanding space. In the first scenario, the energy transfer happens through a specific type of resonance, like a child on a swing being pushed at just the right moment to go higher. In this case, the researcher found that the disorder increased, but only modestly. The energy remained mostly in the original field, with only a small fraction transferring to the new particles. The chaos grew by a factor of about two and a half before the process shut itself off. In the second scenario, the dynamics were far more violent and nonlinear. Here, the energy cascaded wildly, creating a turbulent bath where the disorder increased by more than a thousand times. In this case, the energy was shared almost equally between the original field and the new particles, creating a highly chaotic state that lasted much longer.

The third scenario involved a sudden instability, where the universe's field rolled down a hill from an unstable peak, causing a rapid, exponential growth in the number of particles. In this case, the researcher was able to calculate the disorder analytically, finding that it grew at a steady, linear rate until the field settled into a stable state. By comparing these three distinct behaviors, they demonstrated that the amount of disorder created is not just a number; it acts as a fingerprint that reveals exactly which physical mechanism was at work. They could clearly distinguish between a process driven by resonance, one driven by turbulence, and one driven by a sudden instability, simply by measuring how the disorder evolved over time.

Beyond just measuring the disorder, the researcher asked a deeper question: does this creation of chaos actually push the universe to expand faster? They applied a theoretical framework that suggests the production of disorder can generate a force that accelerates the expansion of space. Their simulations revealed a surprising truth: the acceleration depends not on the amount of disorder itself, but on how much energy is actually transferred into particles. In the violent, turbulent scenarios where a large fraction of the original energy was dumped into new particles, the universe experienced a significant, brief burst of accelerated expansion. However, in the resonant scenario, even though disorder was created, very little energy was transferred, and the universe felt no extra push. This finding clarifies that the "second burst" of acceleration is driven by the heat of the energy transfer, not the chaos of the particles.

The study also looked at what happens after this initial explosion, during the slower, gradual decay of the remaining energy. While this later phase produces the vast majority of the universe's total disorder, it happens so slowly and smoothly that it does not generate any noticeable acceleration. The universe simply expands and cools in the standard way. The researcher concluded that the dramatic, explosive phase of preheating is the only time when the creation of disorder is intense enough to momentarily alter the expansion of the cosmos. Their work provides a clear, physical map of how the early universe transitioned from a smooth, coherent state to the chaotic, hot soup of particles that eventually formed the stars and galaxies we see today, showing that the story of the universe's expansion is written not just in its energy, but in the very nature of its disorder.

Technical Summary: Entropy Production During Reheating and General Relativistic Entropic Acceleration

Problem Statement
Reheating serves as the causal bridge between the end of inflation and the hot Big Bang, transferring energy from the coherent inflaton condensate to relativistic degrees of freedom. While the energetics, field spectra, and gravitational wave signatures of preheating have been extensively studied, the entropy generated during the transition from a coherent state to a broad distribution of quanta remains an underutilized diagnostic. This paper addresses the need to quantify the intrinsic irreversibility of preheating processes. Specifically, it seeks to compute the coarse-grained phase-space entropy directly from nonlinear lattice simulations to distinguish between different reheating regimes (resonance, turbulence, and spinodal instability) and to investigate whether this entropy production can source a "second burst" of accelerated expansion via the General Relativistic Entropic Acceleration (GREA) framework.

Methodology
The author utilizes CosmoLattice, a modern C++ framework for lattice simulations of interacting scalar fields in an expanding universe, to perform classical-statistical simulations. The methodology involves:

Entropy Reconstruction: For each field, the mode occupation numbers $n_k$ are reconstructed from the field and its conjugate momentum on the lattice. The coarse-grained von Neumann (phase-space) entropy is calculated shell-by-shell in momentum space using the formula:
$S(t) = \sum_k \left[ (1 + n_k) \ln(1 + n_k) - n_k \ln n_k \right]$
The study focuses on the produced entropy $\Delta S(t) = S(t) - S(0)$ to subtract the static contribution of initial vacuum fluctuations (the "ultraviolet vacuum floor").
Model Selection: Three canonical single-field models coupled to a daughter scalar $\chi$ via $g^2\phi^2\chi^2$ are simulated:
- Quadratic ( $m^2\phi^2$ ): Harmonic oscillations, matter-like background, driven by broad parametric resonance ( $q_{res}=50$ ).
- Quartic ( $\lambda\phi^4$ ): Anharmonic oscillations, radiation-like background, nearly conformal dynamics ( $q=100$ ).
- Symmetry-Breaking (Tachyonic): Field rolls off an unstable maximum, undergoing spinodal instability ( $q=2$ ).
GREA Implementation: The computed entropy production is coupled back into the cosmic expansion. The standard Friedmann equations are modified to include an entropic force term derived from the second law of thermodynamics ( $T dS = d(\rho a^3) + p d(a^3)$ ). The modified acceleration equation includes a term proportional to the heat transfer rate $T\dot{S}$ :
$\frac{\ddot{a}}{a} = -\frac{4\pi G}{3}(\rho + 3p) + \frac{4\pi G}{3} \frac{T\dot{S}}{a^3 H}$
The simulations track the deceleration parameter $q_{GREA}$ to determine if the entropic force induces acceleration.

Key Results

Quadratic Model ( $m^2\phi^2$ ): Broad parametric resonance drives daughter occupation numbers to $n_k \simeq 7.5 \times 10^2$ in a narrow infrared band. The produced entropy $\Delta S$ grows by a factor of $\simeq 2.5$ and saturates once backreaction shuts off the resonance. Crucially, the entropy resides almost entirely in the daughter sector; the inflaton contribution is negligible or slightly negative (an artifact of vacuum subtraction). Despite significant entropy production, the energy transfer to particles is minimal ( $\rho_{fluc}/\rho \sim 10^{-11}$ ), resulting in a negligible entropic force ( $R \sim 10^{-10}$ ) and no appreciable acceleration.
Quartic Model ( $\lambda\phi^4$ ): The dynamics are strongly nonlinear and radiation-like. Occupation numbers cascade to the infrared, exceeding $n_k \sim 10^{11}$ . The produced entropy grows by over three orders of magnitude ( $\Delta S \simeq 4 \times 10^6$ per comoving box). Rescattering distributes entropy democratically between the inflaton and daughter fields. Because a substantial fraction of coherent energy ( $O(1)$ ) fragments into the turbulent bath, the entropic ratio $R$ reaches $\sim 10$ , driving a substantial extra expansion ( $\Delta N \sim O(1)$ ).
Tachyonic Model (Symmetry-Breaking): The entropy production is analytically tractable. The spinodal growth of long-wavelength modes leads to an exponential increase in occupation numbers, resulting in a strictly linear growth of entropy in time: $S_\phi(t) \simeq (m^4/16\pi)V t$ . This growth saturates abruptly at symmetry breaking ( $t \simeq t_*$ ). Lattice simulations confirm the analytical prediction, reproducing the tanh background profile and the linear entropy law. The conversion of false-vacuum energy into particles is explosive ( $R \sim 10^3$ ), generating a brief but intense burst of entropic acceleration ( $q_{GREA} \ll 0$ ).
GREA Dynamics: The study demonstrates that the entropic acceleration is controlled by the heat (energy transferred to particles, $T\dot{S}$ ), not the phase-space entropy itself.
- In the quadratic case, high entropy but low energy transfer yields no acceleration.
- In the quartic and tachyonic cases, high energy transfer yields significant acceleration.
- The subsequent perturbative reheating phase, while generating the bulk of cosmic entropy ( $S_0 \sim 10^{88}$ ), is quasi-adiabatic ( $R \lesssim 1$ ) and drives no appreciable acceleration.

Significance and Claims
The paper establishes the phase-space entropy as a robust, single-number diagnostic that cleanly separates the resonance-dominated, turbulence-dominated, and spinodal regimes of preheating. It provides a first-principles, lattice-based validation of the GREA framework, confirming that a "second burst" of accelerated expansion can occur during preheating if the process involves a rapid, non-adiabatic transfer of energy into particles.

The author clarifies that while perturbative reheating generates the vast majority of the universe's final entropy, it does so slowly and adiabatically, preventing it from sourcing significant acceleration. Conversely, the violent, non-adiabatic nature of tachyonic and turbulent preheating allows for a transient burst of entropic acceleration. The paper concludes that the entropic force is a transient phenomenon tied to the rate of entropy production relative to expansion, ceasing once thermal equilibrium is established and the comoving entropy freezes.

Technical Summary: Entropy Production During Reheating and General Relativistic Entropic Acceleration

More like this