Cancellation of one-loop correction to soft tensor power spectrum

Basic Information

• Paper ID: 2506.15780
• Title: Cancellation of one-loop correction to soft tensor power spectrum
• Authors: Yohei Ema, Muzi Hong, Ryusuke Jinno, Kyohei Mukaida
• Institutions: CERN, University of Tokyo, Kobe University, KEK
• Classification: astro-ph.CO (Cosmology and Non-Galactic Astrophysics), hep-ph, hep-th
• Publication Date: June 2025 (arXiv v2: November 23, 2025)
• Paper Link: https://arxiv.org/abs/2506.15780

Abstract

This paper demonstrates that on superhorizon scales, the one-loop quantum corrections to tensor perturbations from small-scale (potentially enhanced) scalar perturbations contain no scale-invariant contributions, independent of the specific time evolution details of the inflationary background. The authors derive an effective field theory for soft tensor modes at leading order in gradient expansion by integrating out small-scale scalar fluctuations on the Schwinger-Keldysh path integral contour, i.e., performing loop calculations in the soft external momentum limit. The absence of scale-invariant corrections originates from the diffeomorphism invariance of general relativity and is therefore inevitable.

Research Background and Motivation

Core Problem

The central question addressed in this paper is: Do small-scale enhanced scalar perturbations produce scale-invariant contributions to large-scale tensor perturbations (primordial gravitational waves) through one-loop quantum corrections?

Importance of the Problem

1. Primordial Black Hole (PBH) Formation: If enhanced curvature perturbations exist at small scales, these perturbations may collapse to form primordial black holes after horizon reentry, representing a candidate explanation for dark matter and LIGO/Virgo gravitational wave events.
2. Induced Gravitational Waves: Enhanced scalar perturbations generate induced gravitational waves through second-order effects, detectable by future gravitational wave detectors (e.g., LISA, Einstein Telescope).
3. Theoretical Consistency: Recent studies (particularly Refs. 21-24) claim that in ultra-slow-roll (USR) inflation models, small-scale perturbations produce significant one-loop corrections to large-scale perturbations, contradicting causality expectations (k³ suppression).

Limitations of Existing Approaches

1. Earlier Conclusions: Previous research (Refs. 25-29) demonstrated that in single-clock inflation models, loop corrections from small-scale perturbations to large-scale scalar perturbations are absent.
2. Controversial Claims: Refs. 23,24 assert the existence of scale-invariant corrections from small-scale scalar to large-scale tensor perturbations, challenging conventional understanding.
3. Methodological Issues: Some previous calculations artificially modified mode functions to simulate small-scale enhancement but failed to maintain consistency with background evolution equations.

Research Motivation

The motivation for this work is to use rigorous effective field theory methods and the Schwinger-Keldysh formalism to explicitly demonstrate, without dependence on specific inflationary background evolution details, that one-loop corrections to the soft tensor power spectrum necessarily vanish, thereby resolving this controversy.

Core Contributions

1. Rigorous Proof: Through explicit calculation, demonstrates that in the soft limit (k→0), one-loop corrections from small-scale scalar fields (including potentially enhanced observer fields) to the tensor power spectrum completely vanish.
2. General Result: Proves validity for arbitrary time-dependent background evolution (including transient USR phases), independent of specific inflation models.
3. Effective Field Theory Framework: Derives an effective field theory for soft tensor modes by integrating out small-scale scalar fields on the Schwinger-Keldysh contour, using the Keldysh r/a basis to make causal structure explicit.
4. Key Identity: Discovers and utilizes a crucial identity involving Green's functions (Eq. 3.9): $\frac{\partial}{\partial \log l} G_{rr}(l;\tau_1,\tau_2) = -2il^2 \int d\tau f^2(\tau)c_s^2(\tau)[G_{rr}(l;\tau_1,\tau)G_{ra}(l;\tau_2,\tau) + G_{ra}(l;\tau_1,\tau)G_{rr}(l;\tau_2,\tau)]$ This identity holds for arbitrary background evolution and is central to the cancellation mechanism.
5. Symmetry Explanation: Explains the cancellation mechanism from the perspective of diffeomorphism invariance: the effective Lagrangian forbids $h^a_{ij}h^r_{ij}$ terms, which would produce scale-invariant corrections.
6. Literature Clarification: Points out that the errors in Refs. 23,24 stem from inconsistently modifying mode functions without correspondingly changing background evolution.

Detailed Methodology

Task Definition

Input:

• Action for observer scalar field χ with time-dependent sound speed $c_s(\tau)$, mass $m_χ(\tau)$, and overall coefficient $f(\tau)$
• Graviton tensor perturbations $h_{ij}$

Output:

• One-loop correction to tensor power spectrum in the soft limit (external momentum k→0)

Constraints:

• Maintain general covariance
• Use Bunch-Davies vacuum initial conditions
• Perform calculations on conformal time τ∈-∞,0

Theoretical Framework

1. In-in Formalism and Keldysh r/a Basis

In-in Formalism: Cosmological correlation functions are computed at the end of inflation (τ=0) using Schwinger-Keldysh path integrals with forward (+) and backward (-) time contours.

Field Transformation: From ± basis to r/a basis: $\phi_r = \frac{1}{2}(\phi_+ + \phi_-), \quad \phi_a = \phi_+ - \phi_-$

Propagator Structure:

• Statistical propagator: $G_{rr}(x,y) = \frac{1}{2}[G_>(x,y) + G_<(x,y)]$ (no time-ordering θ function)
• Retarded propagator: $G_{ra}(x,y) = \theta(x^0-y^0)[G_>(x,y) - G_<(x,y)]$
• Important property: $G_{aa} \equiv 0$ (pure a-type operator correlations vanish identically)

Causal Arrow Rules:

• Causal time flow emanates from a fields and enters r fields
• Each vertex contains an odd number of a fields
• Diagrams forming closed causal time loops vanish identically

2. Observer Field Action

Consider the most general observer scalar field: $S = \frac{1}{2}\int d^{d-1}x d\tau \, f^2(\tau)\left[\left(\frac{d\chi}{d\tau}\right)^2 - c_s^2(\tau)[e^{-h}]_{ij}(\partial_i\chi)(\partial_j\chi) - m_\chi^2(\tau)\chi^2\right]$

where:

• $f(\tau)$: arbitrary time-dependent function (typically the scale factor, but not limited to this)
• $c_s(\tau)$: time-dependent sound speed
• $m_χ(\tau)$: time-dependent mass
• $h_{ij}$: graviton satisfying transverse traceless conditions

3. Soft Limit Analysis

Graviton bulk-to-boundary propagators in de Sitter spacetime: $\Delta_{rr}(k;0,\tau) = \frac{H^2}{2k^3}[\cos(k\tau) + k\tau\sin(k\tau)]$$\Delta_{ra}(k;0,\tau) = \frac{iH^2}{k^3}[\sin(k\tau) - k\tau\cos(k\tau)]$

Soft limit behavior: $\lim_{k\to 0}\Delta_{rr}(k;0,\tau) = \frac{H^2}{2k^3}, \quad \lim_{k\to 0}\Delta_{ra}(k;0,\tau) = \frac{iH^2\tau^3}{3}$

Only the statistical propagator $\Delta_{rr}$ contributes $\sim k^{-3}$ terms in the soft limit, so only diagrams with bulk-to-boundary statistical propagators need be considered.

One-loop Calculation

Relevant Feynman Diagrams

One-loop contributions to tensor power spectrum:

←─∥  = ∥─∥ + ∥═∥
1-loop

where:

• First term: three-point vertex contribution (two χ propagators)
• Second term: four-point vertex contribution (one χ propagator)

In-in Interaction Terms

In the r/a basis: $S_{\text{in-in}} = \int d^{d-1}x d\tau \, f^2(\tau)c_s^2(\tau)\left[\frac{1}{2}\hat{h}^a_{ij}(\partial_i\hat{\chi}_r)(\partial_j\hat{\chi}_r) + \hat{h}^r_{ij}(\partial_i\hat{\chi}_r)(\partial_j\hat{\chi}_a) - \frac{1}{2}\hat{h}^a_{ik}\hat{h}^r_{jk}(\partial_i\hat{\chi}_r)(\partial_j\hat{\chi}_r)\right]$

Key Technique: ε-Deformation Method

Consider a deformed quadratic action: $S_{\chi;\epsilon} = \frac{1}{2}\int d^{d-1}x d\tau \, f^2(\tau)\left[\left(\frac{d\chi}{d\tau}\right)^2 - c_s^2(\tau)(1+\epsilon)^2(\partial_i\chi)^2 - m_\chi^2(\tau)\chi^2\right]$

Two calculation approaches:

1. Include $(1+\epsilon)$ at all orders → momentum shift $k \to (1+\epsilon)k$
2. Treat $\epsilon$ perturbatively → diagrammatic expansion

Comparing both methods yields identity (3.9), which is central to the cancellation proof.

Technical Innovations

1. Green's Function Identity: Eq. (3.9) relates diagrams of different topologies (different numbers of propagators), typical of theories with non-trivial symmetries.
2. Path Integral Method: Compared to previous operator formalism-based studies, this work employs Schwinger-Keldysh path integral methods, making calculations more systematic and transparent.
3. Model Independence: By keeping $f(\tau)$, $c_s(\tau)$, $m_χ(\tau)$ arbitrary, the proof demonstrates that cancellation is inevitable whenever mode equations are solved consistently.
4. Integrand-level Cancellation: Demonstrates cancellation at the level of the conformal time integral integrand rather than after integration (avoiding potential contributions from the infinite past).

Experimental Setup

Theoretical Calculation Framework

This is purely theoretical work without numerical experiments or observational data, but provides verifiable predictions:

1. Identity Verification: For any specific inflation model (including transient USR), Green's functions obtained by solving mode equations must satisfy Eq. (3.9).
2. Consistency Check: If any calculation claims scale-invariant one-loop corrections exist, one can verify whether its mode functions satisfy Eq. (3.9).

Calculation Details

1. Dimensionality: Keep spacetime dimension d arbitrary, ultimately taking d=4
2. Regularization: Use dimensional regularization to handle ultraviolet divergences while preserving theory symmetries
3. Boundary Conditions:
   • Initial condition: Bunch-Davies vacuum (τ→-∞)
   • Final state: inflation end surface (τ=0)
4. Angular Integration: Exploit symmetry: $\int \frac{d^{d-1}l}{(2\pi)^{d-1}} \frac{l_il_jl_kl_l}{l^2} = \frac{1}{(d-1)(d+1)}(\delta_{ij}\delta_{kl} + \delta_{ik}\delta_{jl} + \delta_{il}\delta_{jk})\int \frac{d^{d-1}l}{(2\pi)^{d-1}} l^2$

Experimental Results

Main Results

Three-point Vertex Contribution

In the soft limit: $\left\langle \text{three-point} \right\rangle_{\text{soft}} = \frac{i}{4}\int d\tau f^2(\tau)c_s^2(\tau)\int \frac{d^{d-1}l}{(2\pi)^{d-1}} \frac{l_il_jl_kl_l}{l^2}\left[\frac{\partial}{\partial \log l}G_{rr}(l;\tau,\tau)\right]\hat{h}^a_{ij}(\vec{k},\tau)\hat{h}^r_{kl}(-\vec{k})$

Using angular integration and traceless conditions: $= \frac{i}{2(d-1)}\int d\tau f^2(\tau)c_s^2(\tau)\left[\int \frac{d^{d-1}l}{(2\pi)^{d-1}} l^2 G_{rr}(l;\tau,\tau)\right]\hat{h}^a_{ij}(\vec{k},\tau)\hat{h}^r_{ij}(-\vec{k})$

Four-point Vertex Contribution

$\left\langle \text{four-point} \right\rangle_{\text{soft}} = -\frac{i}{2(d-1)}\int d\tau f^2(\tau)c_s^2(\tau)\left[\int \frac{d^{d-1}l}{(2\pi)^{d-1}} l^2 G_{rr}(l;\tau,\tau)\right]\hat{h}^a_{ij}(\vec{k},\tau)\hat{h}^r_{ij}(-\vec{k})$

Total Contribution

$\boxed{\left\langle \text{three-point} + \text{four-point} \right\rangle_{\text{soft}} = 0}$

Conclusion: In the soft limit, one-loop corrections completely vanish with no scale-invariant contributions.

Cancellation Mechanism Analysis

1. Topological Transformation: Eq. (3.9) simplifies the three-point diagram in the soft limit to the same topological structure as the four-point diagram
2. Relative Coefficients: The relative coefficient between the two diagrams is exactly -2:1, leading to complete cancellation
3. Universality: Cancellation is independent of specific forms of $f(\tau)$, $c_s(\tau)$, $m_χ(\tau)$, depending only on consistent mode equation solutions

Literature Clarification

Problems with Refs. 23,24:

• These papers artificially modified mode functions to simulate small-scale enhancement
• But failed to correspondingly modify $f(\tau)$, $c_s(\tau)$, or $m_χ(\tau)$
• Result: Their mode functions do not satisfy Eq. (3.9)
• Therefore, the "scale-invariant corrections" they obtain are artifacts of inconsistent calculations

Verification: The authors checked Eq. (16) of Ref. 23 and confirmed it does not satisfy Eq. (3.9) when used with standard inflation background $a=-1/H\tau$.

Related Work

Loop Corrections to Scalar Perturbations

1. Early Work (Refs. 25-29):
   • Senatore & Zaldarriaga (2010): Prove ζ conservation at all loop orders in single-clock inflation
   • Assassi, Baumann & Green (2013): Understand from symmetry and Ward-Takahashi identities
   • Tanaka & Urakawa (2016): Conservation under heavy field radiation corrections
2. Recent Controversy (2023-2025):
   • Kristiano & Yokoyama (2024): Claim significant one-loop corrections in USR inflation
   • Sparked extensive subsequent research (Refs. 30-56)
   • Most studies support cancellation conclusion (Refs. 40,42,45,46,49,51-55)

Tensor Perturbation Studies

• Limited Research: Compared to scalar perturbations, loop correction studies for tensor perturbations are relatively sparse
• Ref. 57: Claims small but non-zero corrections for superhorizon tensor modes in USR (this paper proves they actually vanish)
• This Work's Contribution: First rigorous proof of one-loop tensor power spectrum correction vanishing

Methodology Comparison

Connection to Symmetries

1. Diffeomorphism Invariance: Core symmetry principle of this paper
2. Soft de Sitter EFT (Refs. 89,90): Forbids $h^a_{ij}h^r_{ij}$ terms from effective field theory perspective
3. Ward-Takahashi Identities: Cancellation of different topology diagrams is typical of gauge theories

Conclusions and Discussion

Main Conclusions

1. Zero-Result Theorem: In the soft limit (k→0), one-loop corrections from small-scale scalar perturbations to large-scale tensor perturbations are strictly zero, vanishing at the integrand level of conformal time integration.
2. Universality: Conclusion holds for arbitrary inflationary background evolution, including:
   • Standard slow-roll inflation
   • Transient ultra-slow-roll (USR) phases
   • Arbitrary evolution of $f(\tau)$, $c_s(\tau)$, $m_χ(\tau)$
3. Physical Origin: Cancellation originates from diffeomorphism invariance of general relativity and is therefore inevitable.
4. Effective Field Theory Interpretation: The soft tensor EFT forbids $h^a_{ij}h^r_{ij}$ terms, while allowed $h^a_{ij}h^a_{ij}$ terms correspond to induced gravitational waves with causality-required k³ suppression.

Limitations

1. Integrand Level: This work focuses on $O(k^{-3})$ terms in the conformal time integral integrand; subleading terms may produce $O(k^{-3})$ contributions after integration, but these:
   • Originate from the infinite past (τ→-∞)
   • Are insensitive to intermediate timescale dynamics (e.g., USR phase)
   • Are expected to be independent of specific USR models
2. One-loop Order: Explicit calculations only reach one-loop, though symmetry arguments apply to all loop orders.
3. Bunch-Davies Vacuum: Calculations assume Bunch-Davies initial conditions; non-BD vacua may exhibit collapse singularities (Refs. 75-78).
4. Observer Field Setup: Focuses on observer scalar fields, though arguments can be generalized to inflaton-driven small-scale perturbations.

Future Directions

1. Multi-loop Generalization: Symmetry arguments suggest cancellation should hold at all loop orders; explicit verification needed.
2. Higher-order Correlators: Extend to three-point, four-point and higher-order correlators of tensor modes.
3. Ward-Takahashi Identities: Explicitly establish connection between Eq. (3.9) and Ward-Takahashi identities (authors promise discussion in Ref. 88).
4. Soft Scalar EFT: Apply the same method to scalar perturbations (in progress, Ref. 88).
5. Inflaton Case: Consider small-scale perturbations driven by inflaton field (rather than observer field).
6. Subleading Term Analysis: Explicitly calculate integral contributions of subleading terms, verifying they depend only on infinite past (as in Ref. 66's logarithmic corrections in de Sitter).

In-depth Evaluation

Strengths

1. Methodological Rigor

• Path Integral Framework: Uses Schwinger-Keldysh formalism, more systematic than operator methods
• Keldysh r/a Basis: Makes causal structure explicit, avoiding time-ordering confusion
• ε-Deformation Technique: Elegantly derives Green's function identity (Eq. 3.9)

2. Universality

• Model Independence: Independent of specific inflation models, applicable to arbitrary background evolution including USR
• Parameter Freedom: $f(\tau)$, $c_s(\tau)$, $m_χ(\tau)$ remain arbitrary, covering broad enhancement mechanisms
• Arbitrary Dimension: Maintains spacetime dimension d as arbitrary until the end

3. Physical Insights

• Symmetry Origin: Traces cancellation to fundamental principle of diffeomorphism invariance
• EFT Perspective: Understands forbidden operator structures from effective field theory viewpoint
• Causality: Consistent with physical causality (k³ suppression)

4. Controversy Resolution

• Clear Error Identification: Clearly explains problems in Refs. 23,24 (mode functions inconsistent with background)
• Operational Test: Provides actionable verification method (verify Eq. 3.9)

5. Clear Presentation

• Step-by-step Derivation: From basic in-in formalism to final result, logically clear
• Diagrammatic Illustration: Uses Feynman diagrams and causal arrows for intuitive presentation
• Technical Details: Sufficient mathematical derivations and physical explanations

Weaknesses

1. Calculation Scope

• One-loop Only: While symmetry arguments hint at multi-loop cancellation, explicit verification lacking
• Integrand Level: Acknowledges subleading terms may contribute after integration, but lacks detailed analysis
• Specific Vacuum: Limited to Bunch-Davies vacuum; other vacuum states not discussed

2. Physical Scenarios

• Observer Fields: Primarily considers observer scalar fields; inflaton case requires further study
• Tensor Modes: Only considers tensor perturbations; complete scalar analysis in preparation

3. Numerical Verification

• No Numerical Checks: As purely theoretical work, lacks numerical verification for specific models
• Observational Connection: Doesn't discuss implications for actual gravitational wave observations

4. Ward-Takahashi Identities

• Unclear Connection: While mentioning connection to Ward-Takahashi identities, detailed discussion deferred to future work

5. Technical Assumptions

• Integration by Parts: Boundary term handling depends on regularization scheme (dimensional regularization)
• Soft Limit Order: Takes soft limit before integration; may differ subtly from integrating first then taking soft limit

Impact

Contribution to Field

1. Controversy Resolution: Definitively refutes recent claims of scale-invariant one-loop corrections in USR inflation
2. Methodology: Demonstrates powerful application of Schwinger-Keldysh EFT methods in cosmology
3. Theoretical Foundation: Reinforces consistency and causality of primordial perturbation theory

Practical Value

1. PBH Constraints: Confirms small-scale enhancement doesn't feedback to large-scale tensor perturbations through one-loop effects
2. Gravitational Wave Detection: Provides reliable theoretical predictions for future experiments (LISA, etc.)
3. Model Building: Guides principles for constructing consistent inflation models

Reproducibility

• Clear Derivation: Calculation steps are transparent and easily reproducible
• Consistency Check: Provides explicit verification method (Eq. 3.9)
• Open Questions: Points to multiple verifiable extensions

Applicable Scenarios

Direct Application

1. Any Single-field Inflation: Including slow-roll, fast-roll, USR and other background evolutions
2. Observer Field Models: Axion curvaton, observer fields, etc.
3. PBH Formation Theory: Various mechanisms involving small-scale enhancement

Generalization Potential

1. Multi-field Inflation: Methods generalizable to multi-field scenarios
2. Non-standard Vacua: Can investigate non-BD vacuum effects
3. Scalar Perturbations: Same framework applicable to scalar modes

Inapplicable Scenarios

1. Non-perturbative Regime: Perturbative expansion fails for very large perturbations
2. Strong Coupling: Requires non-perturbative methods
3. Quantum Gravity Effects: Beyond effective field theory scope

Comparison with Contemporary Work

This paper aligns with Refs. 40,42,45,46,49,51-55 supporting cancellation conclusions, but offers unique advantages:

Theoretical Significance

1. Causality Protection: Confirms quantum corrections don't violate causality (superhorizon conservation)
2. Symmetry Power: Demonstrates how diffeomorphism invariance constrains quantum corrections
3. EFT Foundation: Provides solid EFT basis for cosmological perturbation theory
4. Non-equilibrium Field Theory: Showcases Schwinger-Keldysh method's promise in cosmology

References

Core References

1. In-in Formalism:
   • 58-61 Schwinger, Bakshi & Mahanthappa, Keldysh: Original literature
   • 62-64 Modern reviews: Caron-Huot, Laine & Vuorinen et al.
2. Loop Corrections in Single-clock Inflation:
   • 25 Senatore & Zaldarriaga (2010): Foundational work
   • 28 Assassi, Baumann & Green (2013): Symmetry perspective
   • 29 Tanaka & Urakawa (2016): Conservation theorem
3. Recent Controversy:
   • 21,22 Kristiano & Yokoyama (2024): Claims of corrections
   • 23,24 Ota, Sasaki & Wang (2023): Tensor perturbation claims
   • 40,42,45,46,49,51-55 Multiple studies: Support cancellation conclusion
4. Methodology:
   • 65 Ema & Mukaida (2024): In-in cutting rules
   • 89,90 Cohen & Green, Green & Gupta: Soft de Sitter EFT
   • 77 Arkani-Hamed et al. (2020): Cosmological bootstrap
5. Primordial Black Holes:
   • 9-14 PBH reviews and constraints
   • 67 Kawasaki et al. (2013): Axion curvaton model
   • 68-72 USR inflation-related work

Summary: This is a technically rigorous and physically insightful theoretical work that elegantly resolves the controversy over one-loop corrections to primordial gravitational waves through sophisticated path integral methods, providing a solid theoretical foundation for inflation cosmology and primordial black hole research. Its methodology has broad generalization potential and is expected to have lasting impact on the development of cosmological perturbation theory.