Picard-Fuchs Equations of Twisted Differential forms associated to Feynman Integrals

Pierre Vanhove · 2026

Read Picard-Fuchs Equations of Twisted Differential forms associated to Feynman Integrals by Pierre Vanhove (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-09129.txt.

Introduction

Vanhove extends an old pole-reduction method to find the equations that govern twisted forms tied to Feynman integrals, which encode particle scattering. Regulated versions of these integrals are twisted period integrals, and the extended Griffiths-Dwork method yields their equations. The paper shows the method at work on hypergeometric, elliptic, and Calabi-Yau motives from families of Feynman integrals. It comes from the Regulators V meeting in Pisa.

Catalog record

CREATORPierre Vanhove
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE47986
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.09129

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Picard-Fuchs Equations of Twisted Differential forms associated to Feynman Integrals Thanks: This text is a contribution to the proceedings of the conference Regulators V, 3-13 juin 2024 Department of Mathematics, University of Pisa, Italy. Pierre Vanhove Abstract . Dimensionally or analytically regulated Feynman integrals lead to relative twisted period integrals. We present a recent extension of the Griffiths-Dwork pole reduction algorithm for deriving the D-module of differential operators acting on the twisted differential forms from Feynman integrals. We illustrate the application of this algorithm by providing twisted Picard-Fuchs operators for hypergeometric, elliptic and Calabi-Yau differential motives arising from families of Feynman integrals. 2020 Mathematics Subject Classification Primary 54C40, 14E20; Secondary 46E25, 20C20 Part I Feynman integrals in Physics Feynman integrals are a cornerstone in understanding fundamental interactions and the elementary building blocks of nature. Scattering amplitudes are used in particle physics to compare theoretical predictions with experimental measurements in particle colliders (see [ 1 , 2 , 3 ] for instance), more recently to gravitational wave physics [ 4 , 5 , 6 ] , or the evaluation of the correlation functions of quantum fields at the end of inflation as they provide tools for analysing the formation of structure in the Universe [ 7 , 8 ] . Their accurate calculation, whether analytically or numerically, is needed for precision physics, but this remains a significant hurdle. There are growing evidences [ 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 ] that Feynman integrals needed for precision physics correspond to (relative) period integrals of (singular) Calabi–Yau geometries. Given a family of Feynman integrals attached to a Feynman graph, it is desirable to answer the questions: (1) What is the class of functions to which belongs a given Feynman integral. (2) Determine the complete set of partial differential operators acting on a given family Feynman integrals. Feynman integrals satisfy several remarkable important properties: (1) They are D-finite functions, that is they satisfy a differentiable module of partial differential equations with respect to their parameters [ 27 , 28 ] . (2) Feynman integrals are relative period integrals of a variation of mixed Hodge structures [ 29 , 9 ] . These two properties set the question of analysing the nature of Feynman integral in a clear Hodge theoretical framework. This approach has been developed in [ 29 , 30 , 9 , 31 , 32 , 10 , 33 , 34 , 11 , 35 , 36 , 17 , 37 ] . We recall, in section II , how Feynman integrals and motives are attached to a graph, and we make explicit the appearance of twisted differential in section. In section III , we explain the extension of the Griffths-Dwork reduction for the specific twisted Feynman differentials detailled in [ 38 ] . In section IV , we make explicit twisted differential operators from Feynman integrals: hypergeometric differential operator in section IV.1 , hyperelliptic and elliptic differential operators in section IV.2 , and Calabi–Yau differential operators in section IV.3 . Part II Hodge structures of Feynman integrals II.1. Feynman Graph polynomials Definition II.1 . A Feynman graph \Gamma is a finite collection of vertices V(\Gamma) , edges E(\Gamma) , and half-edges H(\Gamma) satisfying the usual definitions; edges are adjacent to two vertices, and half-edges are adjacent to a single vertex, and allowing multiple edges between pairs of vertices. We let e(\Gamma)=|E(\Gamma)| the number of edges. To each edge of \Gamma we attach a mass variable m_{e}\in\mathbb{R} and to each half-edge we attach a momentum vector p_{h}\in\mathbb{R}^{1,D-1} in the D -dimensional Minkowski space equipped with a metric of signature (1,D-1) . To each half edge of \Gamma attach a vector p_{h}\in\mathbb{C}^{D} subject to the so-called momentum conservation relation \sum_{h\in H(\Gamma)}p_{h}=0. (II.1) We assume that each vertex of \Gamma has a single outgoing half-edge. Therefore, one may view \Gamma as a graph in the usual sense, allowing multiple edges between vertices. For physical processes these vectors belong to of the D -dimensional Minkowski space \mathbb{R}^{1,D-1} . To simplify notation, we view momenta as being attached to vertices, and write p_{v} instead of p_{h} . Furthermore, we consider only the completely massive case with m_{e}^{2}>0 and all external vectors are of non-zero norm p_{v}\cdot p_{v}\neq 0 . We take m_{e},p_{v} as having complex values. The analytic properties of the Feynman integrals are studied by using analytic continuation in the multi-dimensional complex plane spanned by the number of independent scalar products p_{i}\cdot p_{j} with 1\leq i,j\leq|H(\Gamma)| and the masses m_{i}^{2} with 1\leq i\leq e(\Gamma) . We associate to the graph \Gamma two polynomials which are defined as follows [ 39 , 40 ] . Let \{x_{e}\mid e\in e(\Gamma)\} be variables attached to all edges of \Gamma . A spanning tree of \Gamma is a subgraph {\sf T} of \Gamma which contains all vertices of \Gamma , and so that b_{1}(\mathsf{T})=0 and b_{0}(\mathsf{T})=1 . For each spanning tree \mathsf{T} of \Gamma we attach the monomial x^{\mathsf{T}}=\prod_{e\notin{\mathsf{T}}}x_{e} . The first Symanzik polynomial is the polynomial \mathscr{U}=\sum_{\begin{subarray}{c}\text{Spanning}\\ \text{ trees of }\Gamma\end{subarray}}x^{\mathsf{T}}\,. (II.2) A spanning k -forest of \Gamma is a subgraph \mathsf{F} of \Gamma containing all vertices of \Gamma and so that h_{1}(\mathsf{F})=0 and h_{0}(\mathsf{F})=k . We attach the polynomial x^{\mathsf{F}}=\prod_{e\notin\mathsf{F}}x_{i} to each spanning 2-forest. A 2-forest is a disjoint union of two sub-trees \mathsf{F}=\mathsf{T}_{1}\cup\mathsf{T}_{2} , and we define s_{\mathsf{F}}=\sum_{(v_{1},v_{2})\in\mathsf{F}=\mathsf{T}_{1}\cup\mathsf{T}_{2}}p_{v_{1}}\cdot p_{v_{2}} Where the {}\cdot{} -product is the scalar product on \mathbb{C}^{D} . Then \mathscr{V}(\vec{s},\vec{m};D)=\sum_{\begin{subarray}{c}\text{Spanning}\\ \text{ 2-forests of }\Gamma\end{subarray}}s_{\mathsf{F}}x^{\mathsf{F}},\quad\mathscr{F}(\vec{s},\vec{m};D)=\mathscr{U}\times\left(\sum_{e\in E(\Gamma)}m_{e}^{2}x_{e}\right)-\mathscr{V}(\vec{s},\vec{m};D)\,. (II.3) The polynomial \mathscr{F}(\vec{s},\vec{m};D) is called the second Symanzik polynomial of \Gamma , depends the mass parameters and kinematic invariants, respectively: \vec{m}:=\left\{m_{1}^{2},\dots,m_{e(\Gamma)}^{2}\right\}\in\mathbb{R}_{>0}^{e(\Gamma)},\qquad\vec{s}=\{p_{i}\cdot p_{j},i,j\in v(\Gamma)\}. (II.4) When |v(\Gamma)|>D , not all the scalar products are independent, and the number of independent variables satisfies certain Gram determinant conditions [ 41 ] . The discriminant locus of \mathscr{F} depends on the linear relations between the scalar products in \vec{s} . This polynomial is a homogeneous polynomial of degree L+1 in the variables x_{e} for e\in e(\Gamma) , where L=b_{1}(\Gamma) . This L is often called the loop order of \Gamma . Henceforward, we will write instead \mathscr{F} to simplify our notation. II.2. Feynman integrals in parametric representation Using the two polynomials associated to a graph \Gamma , we define a family of Feynman integrals [ 42 , 40 ] I_{\Gamma}(\underline{z};D,\underline{\nu})=\int_{[0,+\infty[^{e(\Gamma)}}\,{\mathscr{U}^{\nu-(l+1){D\over 2}}\over\mathscr{F}^{\nu-l{D\over 2}}}\,\delta\left(\sum_{i=1}^{e(\Gamma)}x_{i}-1\right)\prod_{i=1}^{e(\Gamma)}x_{i}^{\nu_{i}-1}dx_{i}\,. (II.5) where we have set \underline{\nu}:=(\nu_{1},\dots,\nu_{e(\Gamma)}) and collected the kinematic factors (the internal masses m_{i} and the independent scalar products between the external momenta) into \underline{z}=(\vec{s},\vec{m}) . Since the coordinate scaling (x_{1},\dots,x_{e(\Gamma)})\to\lambda(x_{1},\cdots,x_{e(\Gamma)}) leaves invariant the integrand and the domain of integration, we can rewrite this integral as I_{\Gamma}(\underline{z};D,\underline{\nu})=\int_{\Delta_{e(\Gamma)}}\Omega_{\Gamma}^{D,\underline{\nu}} (II.6) with \Omega_{\Gamma}^{D,\underline{\nu}}:={\mathscr{U}^{\nu-(l+1){D\over 2}}\over\mathscr{F}^{\nu-l{D\over 2}}}\,\prod_{e\in e(\Gamma)}x_{e}^{\nu_{e}-1}\,\Omega_{0} (II.7) with the differential e(\Gamma)-1 -form \Omega_{0}:=\sum_{j=1}^{e(\Gamma)}(-1)^{j-1}x_{j}\,dx_{1}\wedge\cdots\wedge\widehat{dx_{j}}\wedge\cdots\wedge dx_{e(\Gamma)}, (II.8) where \widehat{dx_{j}} means that dx_{j} is omitted. The domain of integration is defined as \Delta_{e(\Gamma)}:=\left\{[x_{1},\dots,x_{e(\Gamma)}]\in\mathbb{P}^{e(\Gamma)-1}|x_{i}\in\mathbb{R},x_{i}\geq 0\right\}. (II.9) II.3. Mixed Hodge structures for Feynman graph integrals We define the vanishing loci for the Symanzik polynomials attached to the graph \Gamma : X_{\Gamma;D}=\{\mathscr{F}(\vec{s},\vec{m};D)=0|x_{i}\in\mathbb{P}^{e(\Gamma)-1}(\mathbb{R})\};\qquad Y_{\Gamma}=\{\mathscr{U}=0|x_{i}\in\mathbb{P}^{e(\Gamma)-1}(\mathbb{R})\}\,. (II.10) Notice that the vanishing locus for \mathscr{F} depends on the space-time dimension D through the linear relations between the external momenta. The integrand of the Feynman

Read the complete text