⚛️ high-energy theory

Separate universe in multifield inflation: a phase-space approach

This paper extends the separate-universe approach to multifield inflation using a Hamiltonian phase-space formalism, demonstrating that the approximation accurately describes cosmological perturbations on super-Hubble scales provided specific wavelength bounds are met, thereby validating its use in stochastic inflation and primordial black hole production scenarios.

Original authors: Julien Grain, Hugo Holland

Published 2026-07-20

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Julien Grain, Hugo Holland

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 the universe as a giant, expanding balloon. In the very first fraction of a second after the Big Bang, this balloon didn't just grow; it inflated at a mind-boggling speed, stretching faster than light itself. This era is called inflation. During this wild expansion, tiny quantum jitters—random, microscopic fluctuations in energy—were stretched out until they became the seeds for everything we see today: stars, galaxies, and even you.

But here's the tricky part: to understand how these seeds grew into galaxies, scientists have to do some heavy math. They usually treat the universe as a smooth, perfect background with tiny ripples on top. However, sometimes the universe gets a bit "spicy." In certain models, especially those involving multiple fields (think of them as different types of invisible energy fluids mixing together), the math gets incredibly messy. To simplify this, physicists use a shortcut called the "separate universe" approach. Imagine chopping the universe into millions of tiny, independent bubbles. Inside each bubble, the universe is perfectly smooth and uniform, but the bubbles themselves might be expanding at slightly different rates. It's like looking at a forest by studying one tree at a time, assuming the wind doesn't blow between them. This shortcut is super useful for predicting rare cosmic events, like the formation of Primordial Black Holes—black holes born not from dying stars, but from the collapse of massive energy clumps in the early universe. But does this "one tree at a time" method actually work when the forest is a chaotic, multi-field jungle? That's the big question.

This paper, written by Julien Grain and Hugo Holland, dives deep into the math to see if the "separate universe" shortcut holds up when we have multiple fields of energy interacting in complex ways. They didn't just guess; they used a rigorous mathematical framework called Hamiltonian mechanics (think of it as a super-detailed accounting system for energy and motion) to compare the full, messy reality of the universe with the simplified "separate universe" version.

The authors found that the shortcut does work, but only under very specific conditions. It's not a free pass to ignore the complexity. For the "separate universe" picture to match the real, wiggly universe, the ripples you are studying must be huge—much larger than the "horizon" (the distance light can travel in a given time) and also larger than the "effective mass" of the fields involved. If the ripples are too small or the fields are too light, the shortcut breaks down, and you can't ignore the interactions between the different bubbles.

One of the most interesting discoveries is about how we choose to "look" at the universe. In physics, you can describe the same reality using different coordinate systems, called gauges. The authors showed that the "separate universe" approach works beautifully if you use a specific way of looking called the uniform-expansion gauge. In this view, the universe's expansion rate is the same everywhere, making the "separate bubbles" idea perfectly valid. However, if you try to use other common ways of looking (like the "unitary gauge" or "spatially-flat gauge"), the shortcut can fail because it misses some subtle, non-local connections between the bubbles.

Crucially, the paper argues that the "separate universe" method isn't a magic wand that works everywhere. It explicitly rules out the idea that you can just apply it blindly to any model of inflation. If you are studying models where the fields twist and turn sharply (like a rollercoaster with sudden drops), or if the fields have very light masses, the "separate universe" approximation might give you the wrong answer. The authors suggest that for these tricky cases, you need to be very careful about which mathematical tools you use. They also highlight that to make the shortcut work in the most complex scenarios, you need to use special "gauge-invariant" variables—mathematical tools that don't change just because you shifted your perspective.

In short, Grain and Holland have mapped out the boundaries of a popular cosmological shortcut. They confirmed that the "separate universe" approach is a powerful tool for understanding the early universe and the birth of black holes, but it has a strict rulebook. It works great when the cosmic waves are long and the fields are heavy, and when you look at the universe in the right way. But if you try to force it into situations where the fields are light or the geometry is twisting, the approximation crumbles. This doesn't mean the shortcut is useless; it just means scientists now know exactly when they can trust it and when they need to do the full, complicated math instead.

Technical Summary: Separate Universe in Multifield Inflation: A Phase-Space Approach

Problem Statement
The separate-universe approach is a foundational approximation in cosmological perturbation theory, treating the universe as a collection of homogeneous and isotropic patches. This framework underpins the $\delta N$ -formalism and stochastic inflation, which are essential for studying primordial black hole (PBH) production and large-scale structure formation where curvature power spectrum enhancements are required. While the validity of this approach has been established for single-field inflation, its application to multifield models remains less rigorous. Multifield inflation, particularly in the context of non-linear sigma models, introduces complex dynamics such as curved field spaces, non-geodesic motion, and strong entropic perturbations. A critical challenge is determining the conditions under which the separate-universe approximation (which discards anisotropic degrees of freedom and gradients) accurately reproduces the full cosmological perturbation theory in these multifield scenarios. Furthermore, the matching between the two approaches is known to be gauge-dependent, and the specific validity conditions for multifield models, particularly regarding the choice of gauge and the definition of covariant variables, require clarification.

Methodology
The authors employ a manifestly covariant Hamiltonian formalism to analyze multifield inflation. The study focuses on non-linear sigma models, where $n$ scalar fields $\phi^I$ interact via a potential $V(\phi^I)$ and a field-space metric $G_{IJ}(\phi^I)$ .

Hamiltonian Formulation: The authors derive the Hamiltonian description of general relativity coupled to these scalar fields. They decompose the phase space into background (homogeneous and isotropic) and perturbation variables.
Covariant Variables: To handle the curved field space correctly, the authors introduce manifestly covariant perturbation variables $(Q^I, P_I)$ . These are defined via a canonical transformation from the naive variables (finite differences $\delta \phi^I, \delta \pi_I$ ) using a generating function. This transformation ensures that the variables transform covariantly under field redefinitions. A key technical step involves fixing an ambiguity parameter $\alpha$ in the definition of the covariant momentum to ensure the transformation remains canonical up to cubic order, finding that $\alpha = 1/3$ is required.
Perturbation Theory vs. Separate Universe: The authors derive the equations of motion for cosmological perturbations in the full theory (including gradients and anisotropic modes) and compare them with the equations derived in the separate-universe picture (where gradients and anisotropic modes are set to zero).
Gauge Analysis: The study systematically compares the two approaches in arbitrary gauges and specific gauge fixings, including the unitary gauge, the spatially-flat gauge, and the uniform-expansion gauge. They utilize gauge-invariant variables (Mukhanov-Sasaki variables) and diffeomorphism-invariant variables to assess the matching conditions.

Key Contributions and Results

Validity Conditions: The paper establishes that the separate-universe approach adequately describes cosmological perturbation theory in multifield inflation provided three conditions are met:
1. The wavelength of the modes considered must be larger than several lower bounds. These bounds depend on the cosmological horizon scale and the inverse of the effective Hamiltonian masses of the fields. The effective masses are determined by the coupling potential, the field-space geometry (Riemann curvature), and the background trajectory. Specifically, the condition requires $(k/aH)^2 \ll \min(|1-\epsilon_1|, |1+\epsilon_1|, \{m_a/H\}^2)$ , where $m_a$ are the effective masses of the adiabatic and entropic modes.
2. Anisotropic degrees of freedom must be set to zero (which is inherent to the separate-universe definition).
3. The Lagrange multipliers (perturbed lapse and shift) must be appropriately matched between the two frameworks.
Gauge Dependence and Matching:
- Unitary Gauge: The authors find that a direct matching in the unitary gauge fails because the shift vector perturbation ( $k\delta N_1$ ) is not $k$ -suppressed in the full theory but is absent in the separate-universe picture. However, by introducing diffeomorphism-invariant variables for the adiabatic mode, they demonstrate that the mismatch can be resolved, allowing the separate-universe picture to be valid in this gauge.
- Spatially-Flat Gauge: Attempting to define the spatially-flat gauge in the separate-universe picture leads to a mismatch in the perturbed lapse function. The authors show that fixing the metric perturbation to zero in the separate-universe context actually corresponds to the uniform-expansion gauge in the full theory, not the spatially-flat gauge.
- Uniform-Expansion Gauge: This gauge is shown to be the most natural for the separate-universe approach. The gauge condition $\delta \gamma_1 = 0$ in the separate-universe picture matches the large-scale limit of the uniform-expansion gauge in full perturbation theory. In this gauge, the remaining gauge degree of freedom decouples at large scales, making the approximation robust.
Covariant Momentum Ambiguity: The study highlights a subtle ambiguity in defining covariant momentum variables in the Hamiltonian formalism. While linear order covariance is sufficient for quadratic actions, ensuring a proper canonical transformation at cubic order (necessary for stochastic inflation where background equations of motion cannot be assumed to hold a priori) fixes the momentum definition uniquely with $\alpha = 1/3$ . This choice cancels specific terms involving covariant derivatives of the fields in the cubic Hamiltonian.
Stochastic Inflation Implications: The results refine the understanding of the coarse-graining scale in stochastic inflation. The scale at which modes are separated into "short" (integrated out) and "long" (stochastic) modes must be chosen not only based on the Hubble radius but also relative to the effective masses of the fields. In models with tachyonic instabilities or ultra-slow-roll phases, the effective mass can become very small, potentially breaking the separate-universe approximation or requiring a much stricter coarse-graining scale.

Significance
The paper provides a rigorous, covariant Hamiltonian framework for validating the separate-universe approach in multifield inflation. It clarifies that while the approximation is generally valid at super-Hubble scales, its applicability is contingent on the specific model parameters (field masses and turning rates) and the choice of gauge. By identifying diffeomorphism-invariant variables that decouple from the shift vector mismatch, the authors provide a robust method to apply the separate-universe picture in gauges like the unitary gauge. Furthermore, the work underscores the importance of the uniform-expansion gauge for stochastic $\delta N$ -formalism applications, particularly in scenarios involving primordial black hole production where non-linear effects and strong stochasticity are prevalent. The findings generalize previous single-field results and offer a necessary theoretical foundation for using stochastic inflation in complex, multi-field scenarios.

More like this