arXiv subjects

David Broadhurst

Publications and source records attributed to David Broadhurst.

25 records · Page 2Linked to original sources

arXiv2014

Algebraic geometry informs perturbative quantum field theory

Single-scale Feynman diagrams yield integrals that are periods, namely projective integrals of rational functions of Schwinger parameters. Algebraic geometry may therefore inform us of the types of number to which these integrals evaluate. We give examples at 3, 4 and 6 loops of massive Feynman diagrams that evaluate to Dirichlet $L$-series of modular forms and examples at 6, 7 and 8 loops of counterterms that evaluate to multiple zeta values or polylogarithms of the sixth root of unity. At 8 loops and beyond, algebraic geometry informs us that polylogs are insufficient for the evaluation of terms in the beta-function of $ϕ^4$ theory. Here, modular forms appear as obstructions to polylogarithmic evaluation.

hep-th↗

arXiv2010

Experimental Mathematics and Mathematical Physics

One of the most effective techniques of experimental mathematics is to compute mathematical entities such as integrals, series or limits to high precision, then attempt to recognize the resulting numerical values. Recently these techniques have been applied with great success to problems in mathematical physics. Notable among these applications are the identification of some key multi-dimensional integrals that arise in Ising theory, quantum field theory and in magnetic spin theory.

math-ph↗

arXiv2010

Feynman's sunshine numbers

This is an expansion of a talk for mathematics and physics students of the Manchester Grammar and Manchester High Schools. It deals with numbers such as the Riemann zeta value zeta(3)=sum_{n>0}1/n^3. Zeta values appear in the description of sunshine and of relics from the Big Bang. They also result from Feynman diagrams, which occur in the quantum field theory of fundamental particles such as photons, electrons and positrons. My talk included 7 reasonably simple problems, for which I here add solutions, with further details of their context.

physics.pop-ph↗

arXiv2010

Dickman polylogarithms and their constants

The Dickman function F(alpha) gives the asymptotic probability that a large integer N has no prime divisor exceeding N^alpha. It is given by a finite sum of generalized polylogarithms defined by the exquisite recursion L_k(alpha)=- int_alpha^{1/k} dx L_{k-1}(x/(1-x))/x with L_0(alpha)=1. The behaviour of these Dickman polylogarithms as alpha tends to 0 defines an intriguing series of constants, C_k. I conjecture that exp(gamma z)/Gamma(1-z) is the generating function for sum_{k\ge0} C_k z^k. I obtain high-precision evaluations of F(1/k), for integers k<11, and compare the Dickman problem with problems in condensed matter physics and quantum field theory.

math-ph↗

arXiv2008

Solutions by radicals at singular values k_N from new class invariants for N \equiv 3 mod 8

For square-free $N\equiv3$ mod 8 and $N$ coprime to 3, I show how to reduce the singular value $k_N$ to radicals, using a novel pair $[f,g]$ of real numbers that are algebraic integers of the Hilbert class field of $Q(\sqrt{-N})$. One is a class invariant of modular level 48, with a growth $g=α(N)\exp(π\sqrt{N}/48)+o(1)$, where $α(N)\in[-\sqrt2,\sqrt2]$ is uniquely determined by the residue of $N$ modulo 64. Hence $g$ is a very economical generator of the class field. For prime $N\equiv3$ mod 4, I conjecture that the Chowla--Selberg formula provides an algebraic {\em unit} of the class field and determine its minimal polynomial for the 155 cases with $N<2000$. For N=2317723, with class number $h(-N)=105$, I compute the minimal polynomial of $g$ in 90 milliseconds. Its height is smaller than the {\em cube} root of the height of the generating polynomial found by the double eta-quotient method of {\em Pari-GP}. I reduce the complete elliptic integral $K_{2317723}$ to radicals and values of the $Γ$ function, by determining the Chowla--Selberg unit and solving the septic, quintic and cubic equations that generate sub-fields of the class field. I conclude that the residue 3 modulo 8, initially discarded in elliptic curve primality proving, outperforms the residue 7.

math-ph↗

arXiv2008

Elliptic integral evaluations of Bessel moments

We record what is known about the closed forms for various Bessel function moments arising in quantum field theory, condensed matter theory and other parts of mathematical physics. More generally, we develop formulae for integrals of products of six or fewer Bessel functions. In consequence, we are able to discover and prove closed forms for $c_{n,k}:=\int_0^\infty t^k K_0^n(t) {\rm d}t$ with integers $n=1,2,3,4$ and $k\ge0$, obtaining new results for the even moments $c_{3,2k}$ and $c_{4,2k}$. We also derive new closed forms for the odd moments $s_{n,2k+1}:=\int_0^\infty t^{2k+1}I_0^{}(t) K_0^{n-1}(t) {\rm d}t$ with $n=3,4$ and for $t_{n,2k+1}:=\int_0^\infty t^{2k+1}I_0^2(t) K_0^{n-2}(t) {\rm d}t$ with $n=5$, relating the latter to Green functions on hexagonal, diamond and cubic lattices. We conjecture the values of $s_{5,2k+1}$, make substantial progress on the evaluation of $c_{5,2k+1}$, $s_{6,2k+1}$ and $t_{6,2k+1}$ and report more limited progress regarding $c_{5,2k}$, $c_{6,2k+1}$ and $c_{6,2k}$. In the process, we obtain 8 conjectural evaluations, each of which has been checked to 1200 decimal places. One of these lies deep in 4- dimensional quantum field theory and two are probably provable by delicate combinatorics. There remains a hard core of five conjectures whose proofs would be most instructive, to mathematicians and physicists alike.

hep-th↗

arXiv2008

Elliptic integral evaluation of a Bessel moment by contour integration of a lattice Green function

A proof is found for the elliptic integral evaluation of the Bessel moment $$M:=\int_0^\infty t I_0^2(t)K_0^2(t)K_0(2t) {\rm d}t ={1/12} {\bf K}(\sin(π/12)){\bf K}(\cos(π/12)) =\frac{Γ^6(\frac13)}{64π^22^{2/3}}$$ resulting from an angular average of a 2-loop 4-point massive Feynman diagram, with one internal mass doubled. This evaluation follows from contour integration of the Green function for a hexagonal lattice, thereby relating $M$ to a linear combination of two more tractable moments, one given by the Green function for a diamond lattice and both evaluated by using W.N. Bailey's reduction of an Appell double series to a product of elliptic integrals. Cubic and sesquiplicate modular transformations of an elliptic integral from the equal-mass Dalitz plot are proven and used extensively. Derivations are given of the sum rules $$\int_0^\infty(I_0(a t)K_0(a t)-\frac{2}π K_0(4a t) K_0(t))K_0(t) {\rm d}t=0$$ with $a>0$, proven by analytic continuation of an identity from Bailey's work, and $$\int_0^\infty t I_0(a t)(I_0^3(a t)K_0(8t)- \frac{1}{4π^2} I_0(t)K_0^3(t)) {\rm d}t=0$$ with $2\ge a\ge0$, proven by showing that a Feynman diagram in two spacetime dimensions generates the enumeration of staircase polygons in four dimensions.

hep-th↗