Hamiltonians to all Orders in Perturbation Theory and Higher Loop Corrections in Single Field Inflation with PBHs Formation

This paper derives all-order interaction Hamiltonians and non-linear field relations for single-field inflation with a transient ultra-slow-roll phase, demonstrating that loop corrections to long-wavelength perturbations grow rapidly as $(\Delta N \mathcal{P}_e L)^L$, thereby causing a breakdown of perturbative control at the fourth loop order in standard models used for primordial black hole formation.

The Gist

The Big Picture: Building a Cosmic Castle

Imagine the early universe as a construction site where a giant cosmic castle (the universe we see today) is being built. The architects use a specific blueprint called "Inflation," a period where the universe expands incredibly fast.

Usually, this construction is smooth and steady. However, the authors of this paper are studying a specific, tricky scenario called Ultra Slow-Roll (USR). In this scenario, the construction crew hits a patch of "super-slippery ice." For a short time, the building materials (quantum fluctuations) start sliding and piling up uncontrollably instead of settling down.

The goal of this paper is to figure out: If we build on this slippery ice, does the pile get so high that the whole structure collapses under its own weight?

The Problem: The "Snowball" Effect

In standard physics, when you calculate how these piles of material interact, you usually look at the big, obvious interactions first. But in this "slippery ice" scenario, the authors found that the tiny, hidden interactions (called loop corrections) start to act like a runaway snowball.

Think of it like this:

• The Main Event: A few large boulders (the main inflation) rolling down a hill.
• The Loops: Tiny pebbles bouncing off the boulders.
• The Issue: On normal ground, the pebbles just bounce off and stop. But on this "slippery ice," every time a pebble bounces, it gains a little more energy and creates more pebbles. If you keep calculating these bounces (loops), the number of pebbles explodes.

The paper asks: At what point does the pile of pebbles become so massive that our math breaks down?

The Tool: The "Master Key" (EFT)

Calculating these interactions is usually a nightmare. It's like trying to solve a puzzle where every piece changes shape as you touch it. The authors used a special tool called Effective Field Theory (EFT).

Think of EFT as a Master Key or a Universal Translator. Instead of trying to solve the puzzle piece by piece (calculating every single interaction one by one), they found a single, compact formula that describes all the interactions at once, no matter how complex they get.

• They translated the messy, complicated math of gravity into a simpler language involving a "Goldstone field" (let's call it $\pi$).
• They created a dictionary to translate this simple language back into the language of the universe's shape (curvature perturbations, or $R$).

This allowed them to see the whole picture without getting lost in the details of every single brick.

The Discovery: The "Sharp Turn" Trap

The authors focused on a specific setup used to explain Primordial Black Holes (PBHs). These are tiny black holes formed in the early universe, which some scientists think could be the "Dark Matter" holding galaxies together.

To make these black holes, the universe needs to pause on that "slippery ice" for a specific amount of time (about 2.5 "e-folds," or expansion cycles) to pile up enough material. Then, it must snap back to normal speed instantly.

Here is the critical finding:

1. The Sharp Turn: The transition from the "slippery ice" back to "normal ground" is like a car hitting a brick wall instead of a gentle ramp.
2. The Explosion: Because the turn is so sharp, the tiny pebbles (loop corrections) don't just bounce; they scream. The math shows that for every extra layer of calculation (loop) you add, the error grows by a massive factor.
3. The Breaking Point: The authors calculated that if you try to build this specific type of black hole, the math stays under control for the first few layers of calculation. But by the 4th layer (4 loops), the numbers become so huge that the theory loses control. It's like trying to stack a tower of Jenga blocks where every new block is 10 times heavier than the one below it; the tower collapses before you finish the 4th floor.

The Two Sources of Chaos

The authors broke the chaos down into two sources:

1. The Bulk (The Ice Patch): While the universe is sliding on the ice, the errors grow, but they grow slowly enough that you can actually sum them all up into a neat, final answer. It's like a slow leak in a boat; you can patch it.
2. The Boundary (The Wall): The moment the universe hits the "wall" to snap back to normal speed, that is where the real disaster happens. The sharpness of this wall creates mathematical "spikes" (delta functions). These spikes get worse and worse with every layer of calculation. This is the part that refuses to be tamed and causes the theory to break.

The Conclusion

The paper concludes that the popular model for creating Primordial Black Holes using this "slippery ice" method is mathematically unstable in its simplest form.

• If the transition back to normal is too sharp (instant), the math breaks at the 4th loop.
• The longer the universe stays on the "ice," the faster the math breaks.
• To fix this, the transition would need to be a gentle ramp, not a wall. However, calculating that gentle ramp is so difficult that the authors couldn't do it with their current tools; it would require a supercomputer simulation.

In short: The authors built a universal calculator for cosmic interactions and found that the specific recipe for making "Dark Matter Black Holes" is too volatile. The math explodes before the recipe is finished, suggesting that this specific way of making black holes might not work as simply as we thought.

Technical Summary

Technical Summary: Hamiltonians to All Orders in Perturbation Theory and Higher Loop Corrections in Single Field Inflation with PBHs Formation

Problem Statement
The paper addresses the theoretical challenge of calculating cosmological correlators and loop corrections to arbitrary orders in perturbation theory within single-field inflation models featuring a transient Ultra-Slow-Roll (USR) phase. Such models are frequently employed to generate Primordial Black Holes (PBHs) as dark matter candidates. A critical issue in these setups is the potential breakdown of perturbative control due to large quantum loop corrections. Previous studies, notably by Kristiano and Yokoyama [6] and subsequent works [16, 28–30, 41], have indicated that one-loop corrections from USR modes can significantly affect long-wavelength CMB scale perturbations, potentially destroying the validity of the perturbative expansion. However, a complete analysis of higher-order loops (two-loop and beyond) has been hindered by the technical complexity of deriving interaction Hamiltonians beyond the quartic order and the sheer number of Feynman diagrams involved.

Methodology
The authors employ the Effective Field Theory (EFT) of inflation formalism in the decoupling limit, where gravitational back-reactions (lapse and shift functions) are neglected. This approach allows for a significant simplification in deriving the action and interaction Hamiltonians for the Goldstone field $\pi$, which encodes the breaking of time diffeomorphism invariance.

1. All-Orders Hamiltonian Derivation: The authors derive a compact, non-perturbative expression for the interaction Hamiltonian to all orders in perturbation theory. By expanding the Hubble parameter $H(t+\pi)$ and its derivatives in the matter action, they obtain a closed-form expression for the $n$-th order action. This leads to a non-perturbative interaction Hamiltonian $H_I$ expressed in terms of $\pi$.
2. Non-linear Mapping: A non-linear relation between the Goldstone field $\pi$ and the comoving curvature perturbation $\mathcal{R}$ is established to all orders. This "dictionary" is essential for translating results from the $\pi$ basis (where the Hamiltonian is simple) to the $\mathcal{R}$ basis (where observables are defined).
3. Loop Calculation Strategy: To calculate $L$-loop corrections, the authors focus on a specific subset of Feynman diagrams: those involving a single vertex with $L$ loops attached. This corresponds to an interaction Hamiltonian of order $n = 2L + 2$. The authors argue that for a given $L$, these single-vertex diagrams provide the dominant contribution compared to multi-vertex diagrams, primarily because the singular terms associated with the transition from USR to Slow-Roll (SR) become more pronounced at higher interaction orders.
4. In-In Formalism: The loop corrections are computed using the in-in formalism. The analysis distinguishes between contributions from the "bulk" of the USR phase (where $\eta = -6$ is constant) and the "boundary" at the transition time $\tau_e$ (where $\eta$ experiences a sharp jump, modeled by a Dirac delta function and its derivatives).

Key Contributions and Results

• Non-Perturbative Hamiltonian: The paper provides a general, non-perturbative expression for the interaction Hamiltonian in the decoupling limit (Eq. 3.22) valid for any single-field model with sound speed $c_s=1$. In the bulk of the USR phase, this takes the form $H_I \propto (e^{-\eta H \pi} - 1)\dot{\pi}^2 + (e^{\eta H \pi} - 1)(\partial \pi)^2$.
• All-Orders Relation: A general formula relating $\mathcal{R}$ to $\pi$ to arbitrary orders is derived (Eq. 3.29), enabling the calculation of correlators without truncating the non-linear mapping.
• Scaling of Loop Corrections: The authors calculate the $L$-loop corrections to the power spectrum. They find that the fractional loop correction scales as:
  $$\frac{\Delta P}{P_{cmb}} \sim \left( 18 L \Delta N P_e \right)^L$$
  where $\Delta N$ is the duration of the USR phase, $P_e$ is the peak power spectrum at the end of USR, and $L$ is the loop order.
• Bulk vs. Boundary Contributions:
  • Bulk: The contributions from the bulk of the USR phase can be resummed into a non-perturbative result involving hypergeometric functions (Eq. 5.34). These contributions are found to be convergent for large $L$.
  • Boundary: The contributions from the sharp transition boundary (where $\dot{\eta}$ and higher derivatives are singular) dominate the loop corrections. The dimensionless coefficient $R_b(L)$ associated with the boundary scales as $(18L)^L$ for large $L$. Consequently, the boundary contributions do not converge and grow rapidly with $L$.
• Critical Loop Order ($L_c$): For a given duration $\Delta N$, there exists a critical loop order $L_c$ beyond which the loop corrections exceed the tree-level result, indicating a loss of perturbative control.
  • In the conventional PBH formation setup where $\Delta N \simeq 2.5$ (required to enhance the power spectrum by $\sim 7$ orders of magnitude), the authors find that $L_c \simeq 3.4$.
  • This implies that for $\Delta N \simeq 2.5$, the perturbative expansion breaks down at the four-loop order ($L=4$).

Significance and Claims
The paper claims to provide a powerful framework for calculating cosmological correlators to any order in perturbation theory using the EFT of inflation. The primary significance of the results lies in the demonstration that single-field USR models with sharp transitions, often used for PBH formation, may not be under perturbative control.

The authors conclude that the rapid growth of loop corrections is an inevitable consequence of the sharpness of the transition from the USR phase to the final SR attractor. The singularities introduced by the jump in $\eta$ (modeled as delta functions) drive the divergence of the perturbation series at higher loop orders. While the authors acknowledge that a smooth transition or an extended relaxation period (where $|h| \ll 1$) could potentially tame these corrections, they note that such scenarios are analytically intractable and would require numerical investigation.

The paper does not claim to resolve the issue of PBH formation but rather highlights a theoretical limitation: in the simplest realization of these models (instantaneous sharp transition), the perturbative approach fails at relatively low loop orders ($L=4$). The authors suggest that this behavior may be generic for inflationary models with sharp localized features in the potential. They also note that while renormalization is necessary for finite physical quantities, it is unlikely to cancel the leading-order growth of the loop corrections due to the lack of an underlying symmetry.