Quantum de Sitter Entropy and Sphere Partition Functions: A-Hypergeometric Approach to Higher Loop Corrections

Bhavya Bandaru

2024-11-26

Abstract:In order to find quantum corrections to the de Sitter entropy, a new approach to higher loop Feynman integral computations on the sphere is presented. Arbitrary scalar Feynman integrals on a spherical background are brought into the generalized Euler integral (A-hypergeometric series/GKZ system) form by expressing the massive scalar propagator as a quotient of a bivariate radial Mellin transform of the massless scalar propagator in one higher dimensional Euclidean flat space. This formulation is expanded to include massive and massless vector fields by construction of similar embedding space propagators. Vector Feynman integrals are shown to be sums over generalized Euler integrals formed of underlying scalar Feynman integrals. Granting existence of general spin embedding space propagators, the same is shown to be true for general spin Feynman integrals.

High Energy Physics - Theory

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to calculate the quantum corrections to the entropy of de Sitter space. Specifically, the author proposes a new method to calculate the higher - order loop Feynman integrals, which are carried out on the sphere. Through this method, the author hopes to calculate the quantum corrections to the de Sitter entropy more accurately, so as to better understand the performance of quantum gravity in de Sitter space.

### Core problems of the paper

1. **Importance of quantum corrections**:
- The macroscopic entropy \( S \) of de Sitter space can be defined by the Euclidean path integral, that is, \( S=\log Z \), where \( Z \) is the path integral on the \((d + 1)\)-sphere.
- The entropy \( S_{\text{tree}}=\frac{A_{\text{hor}}}{4G} \) in the tree - level approximation cannot provide more characteristics of de Sitter space and can only be regarded as a dimensionless coupling constant of a low - energy effective field theory.
- Higher - order quantum corrections (non - local quantum corrections) can provide important information about the space geometry, and each higher - order correction contains information about the space geometry.

2. **Methods for calculating higher - order quantum corrections**:
- Traditional calculation methods encounter difficulties in Feynman integrals on the sphere, especially the integral forms of higher - order loop diagrams are complex and difficult to solve analytically.
- The author proposes a new method, which converts any scalar Feynman integral into the form of a generalized Euler integral (A - hypergeometric series/GKZ system), and simplifies the calculation process by embedding the space to represent the mass scalar propagator.
- This method is not only applicable to scalar fields, but also can be extended to vector fields and other spin fields by constructing similar embedding - space propagators.

### Specific technical details

1. **Embedding - space representation of scalar propagators**:
- The mass scalar propagator \( G(\hat{X},\hat{Y}) \) can be represented on the sphere as:
\[
G(\hat{X},\hat{Y})=\frac{\Gamma(\Delta)\Gamma(d-\Delta)}{(4\pi)^{\frac{d + 1}{2}}} {}_2F_1(\Delta,d-\Delta;\frac{d + 1}{2};\frac{1+\hat{X}\cdot\hat{Y}}{2})
\]
- Through the embedding - space representation, the propagator can be written in the momentum - space form:
\[
G(\hat{X},\hat{Y})\propto\int_{0}^{\infty}\frac{d\mu}{\mu}\mu^{\Delta}\int d^{d + 2}P\frac{e^{-iP(\hat{X}-\mu\hat{Y})}}{P^{2}}
\]

2. **Generalized Euler integral form of Feynman integrals**:
- Any scalar Feynman diagram can be directly converted into the generalized Euler integral form through its incidence matrix \( L \):
\[
I_F\propto\int_{0}^{\infty}\left(\prod_{i = 1}^{n}\frac{d\lambda_i}{\lambda_i}\frac{d\mu_i}{\mu_i}\lambda_i^{d-\Delta_i}\mu_i^{\Delta_i}\right)\frac{1}{\left(\det(1 + L^T L)\right)^{\frac{d + 2}{2}}}
\]
- For example, the Feynman integral of the 3 - melon diagram is:
\[
I_{2,3}\propto\int_{0}^{\infty}\left(\prod_{i = 1}^{3}\frac{d\lambda_i}{\lambda_i}\frac{d\mu_i}{\mu_i}\right)