Table of Contents
Fetching ...

The five-twist identity for Feynman periods

Oliver Schnetz

TL;DR

This work introduces a new five-twist identity for Feynman periods, acting on five-vertex cuts of completed primitive graphs, and proves its independence from the known twist, Fourier identity/split, and planar duality in $\\phi^4$ theory. The authors develop a general framework for internally completed four-point integrals using graphical functions and dual planar graphs, showing that double transpositions of external legs yield invariances under suitable degree and planarity conditions. They then derive the five-twist by inserting a planar dual $G_1^*$ and performing a sequence of decomplete–dualize–twist operations, establishing a robust mechanism for relating subgraph insertions without altering the overall period. Applied to $\\phi^4$ theory, the five-twist generates loop-ordered identities up to eleven loops, including several relations between $\\phi^4$ and non-$\\phi^4$ periods and evidence of independence from established identities (e.g., via $P_{9,103}$), suggesting a richer combinatorial structure of periods and potential extensions to broader QFT contexts.

Abstract

We prove a new identity for Feynman periods that acts on five-vertex cuts of completed primitive Feynman graphs. It is shown that in $φ^4$ theory this identity is independent from existing identities which are the twist, the Fourier identity and the Fourier split.

The five-twist identity for Feynman periods

TL;DR

This work introduces a new five-twist identity for Feynman periods, acting on five-vertex cuts of completed primitive graphs, and proves its independence from the known twist, Fourier identity/split, and planar duality in theory. The authors develop a general framework for internally completed four-point integrals using graphical functions and dual planar graphs, showing that double transpositions of external legs yield invariances under suitable degree and planarity conditions. They then derive the five-twist by inserting a planar dual and performing a sequence of decomplete–dualize–twist operations, establishing a robust mechanism for relating subgraph insertions without altering the overall period. Applied to theory, the five-twist generates loop-ordered identities up to eleven loops, including several relations between and non- periods and evidence of independence from established identities (e.g., via ), suggesting a richer combinatorial structure of periods and potential extensions to broader QFT contexts.

Abstract

We prove a new identity for Feynman periods that acts on five-vertex cuts of completed primitive Feynman graphs. It is shown that in theory this identity is independent from existing identities which are the twist, the Fourier identity and the Fourier split.
Paper Structure (4 sections, 7 theorems, 42 equations, 6 figures)

This paper contains 4 sections, 7 theorems, 42 equations, 6 figures.

Key Result

Proposition 2

The Feynman period (PG) exists if and only if edges that cut the completion $\overline{G}$ with at least two vertices on either side always have total weight greater than $D/\lambda$.

Figures (6)

  • Figure 1: The bubble and the tetrahedron are the smallest primitive graphs in $\phi^4$ theory.
  • Figure 2: The five-twist on the completed graph (left) and as reflections along diagonals of a square in the decompleted graph (right). The shaded areas stand for any subgraphs. Only if the graph $G_1$ has specific properties, the five-twist becomes an identity for Feynman periods.
  • Figure 3: The smallest nontrivial five-twist links $P_{7,2}$ to the non-$\phi^4$ period $P^{\mathrm{non}\,\phi^4}_{7,17}$. The dashed edge on the right hand side has weight $-1$ (a numerator edge), all other edges have weight $1$.
  • Figure 4: Examples of insertions $G_1$ in the five-twist.
  • Figure 5: The only four vertex split of the $\phi^4$ graph $P_{9,103}$ is at the gray vertices. The twist at these vertices differs from the five-twist along the subgraph that is depicted at the left of Figure \ref{['fig:G1']}. The vertex $\infty$ is the center of the small square.
  • ...and 1 more figures

Theorems & Definitions (15)

  • Proposition 2
  • Definition 3
  • Proposition 4
  • proof
  • Proposition 5
  • proof
  • Corollary 6
  • proof
  • Remark 7
  • Proposition 8
  • ...and 5 more