Skip to content

DirichletBeta ​

The Dirichlet beta function β(s)=∑n≥0(−1)n(2n+1)−s\beta(s) = \sum_{n\ge0} (-1)^n (2n+1)^{-s}, the L-function of the nontrivial character mod 4 — entire, with β(1)=π/4\beta(1) = \pi/4 and β(2)=G\beta(2) = G (Catalan).

✓ checked against that system · ! checked, with disagreements (hover a chip)WikipediaDirichlet beta functionMathWorldDirichletBetaFunctionWikidataQ1227706Wolfram LanguageDirichletBeta!
DirichletBeta(s)the Dirichlet beta function β(s)\beta(s).
mpmathdirichlet(s, [0, 1, 0, -1])Juliaenumeratio_ball(s -> (polylog(s, onei(CC)) - polylog(s, -onei(CC))) / (2 * onei(CC)), s)
Typed signatures
TypePackageApplies
(complex | infinity) -> numberanalyticreplaces compute-enginealways
(complex | infinity) -> numbercompute-enginealways

Domain: Special functions

Details
  • β(s)=4−s(ζ(s,14)−ζ(s,34))\beta(s) = 4^{-s}\left(\zeta(s, \tfrac14) - \zeta(s, \tfrac34)\right) in terms of HurwitzZeta, which is how it is evaluated numerically (complex ss included).
  • Odd positive integers have closed forms in π\pi and the Euler numbers: β(2k+1)=(−1)kE2k π2k+1/(4k+1(2k)!)\beta(2k+1) = (-1)^k E_{2k}\,\pi^{2k+1} / (4^{k+1}(2k)!) — β(1)=π/4\beta(1) = \pi/4 (Leibniz), β(3)=π3/32\beta(3) = \pi^3/32, β(5)=5π5/1536\beta(5) = 5\pi^5/1536.
  • Even positive integers do not: β(2)=G=0.9159…\beta(2) = G = 0.9159\ldots is Catalan's constant, and β(4),β(6),…\beta(4), \beta(6), \dots stay symbolic.
  • Nonpositive integers are Euler numbers: β(−2k)=E2k/2\beta(-2k) = E_{2k}/2 (β(0)=12\beta(0) = \tfrac12, β(−2)=−12\beta(-2) = -\tfrac12, β(−4)=52\beta(-4) = \tfrac52), and β(−(2k+1))=0\beta(-(2k+1)) = 0.
  • Functional equation β(1−s)=(π/2)−ssin⁡(πs/2) Γ(s) β(s)\beta(1-s) = (\pi/2)^{-s} \sin(\pi s/2)\,\Gamma(s)\,\beta(s); entire, no poles.

Plot

ComplexPlot
real slice

Examples

2 more kept as data, checked against the oracles
Basic #

β(1)=π/4\beta(1) = \pi/4: Leibniz's series 1−13+15−⋯1 - \tfrac13 + \tfrac15 - \cdots

live ↗
InDirichletBeta(1)\mathrm{DirichletBeta}(1)Outπ4\frac{π}{4}

β(2)=G\beta(2) = G, Catalan's constant

live ↗
InDirichletBeta(2)\mathrm{DirichletBeta}(2)OutG⁡\operatorname{G}

β(3)=π3/32\beta(3) = \pi^3/32

live ↗
InDirichletBeta(3)\mathrm{DirichletBeta}(3)Outπ332\frac{π^3}{32}

Numeric for a floating-point argument — through 4−s(ζ(s,14)−ζ(s,34))4^{-s}(\zeta(s,\tfrac14) - \zeta(s,\tfrac34)), so it answers to the engine's working precision rather than a double's

live ↗
InDirichletBeta(2.5)\mathrm{DirichletBeta}(2.5)Out0.948 622 174 037 054 707 4450.948\,622\,174\,037\,054\,707\,445

A symbolic argument stays symbolic

live ↗
InDirichletBeta(s)\mathrm{DirichletBeta}(s)OutDirichletBeta(s)\mathrm{DirichletBeta}(s)
live ↗
InMap(DirichletBeta,[2,3])\mathrm{Map}(\mathrm{DirichletBeta}, \bigl\lbrack2, 3\bigr\rbrack)Out[G⁡,π332]\bigl\lbrack\operatorname{G}, \frac{π^3}{32}\bigr\rbrack
Scope #

Complex arguments, here on the critical line

live ↗
InDirichletBeta(0.5+14i)\mathrm{DirichletBeta}(0.5+14i)Out1.537 115 438 440 316+1.343 451 426 867 757 3i1.537\,115\,438\,440\,316+1.343\,451\,426\,867\,757\,3i

Threads elementwise over an array

live ↗
InDirichletBeta([[1,−1],[−1,1]])\mathrm{DirichletBeta}(\bigl\lbrack\bigl\lbrack1, -1\bigr\rbrack, \bigl\lbrack-1, 1\bigr\rbrack\bigr\rbrack)Out[[π4,0],[0,π4]]\bigl\lbrack\bigl\lbrack\frac{π}{4}, 0\bigr\rbrack, \bigl\lbrack0, \frac{π}{4}\bigr\rbrack\bigr\rbrack

β(s)→1\beta(s) \to 1 as s→∞s \to \infty

live ↗
InDirichletBeta(∞)\mathrm{DirichletBeta}(∞)Out11

Uncertainty propagation: β(2±0.01)=G±0.00082\beta(2 \pm 0.01) = G \pm 0.00082

live ↗
InDirichletBeta(Around(2,0.01))\mathrm{DirichletBeta}(\mathrm{Around}(2, 0.01))OutAround(0.915 965 594 177 219,0.000 815 807 361 165 928)\mathrm{Around}(0.915\,965\,594\,177\,219, 0.000\,815\,807\,361\,165\,928)

To exactly 40 significant digits, as many as asked for

live ↗
InN(DirichletBeta(2),40)N(\mathrm{DirichletBeta}(2), 40)Out0.915 965 594 177 219 015 054 603 514 932 384 110 774 10.915\,965\,594\,177\,219\,015\,054\,603\,514\,932\,384\,110\,774\,1
live ↗
InFullSimplify(MatrixFunction(DirichletBeta,[[1,−1],[−1,1]]))\mathrm{FullSimplify}(\mathrm{MatrixFunction}(\mathrm{DirichletBeta}, \bigl\lbrack\bigl\lbrack1, -1\bigr\rbrack, \bigl\lbrack-1, 1\bigr\rbrack\bigr\rbrack))Out[[14+G⁡2,14−G⁡2],[14−G⁡2,14+G⁡2]]\bigl\lbrack\bigl\lbrack\frac{1}{4}+\frac{\operatorname{G}}{2}, \frac{1}{4}-\frac{\operatorname{G}}{2}\bigr\rbrack, \bigl\lbrack\frac{1}{4}-\frac{\operatorname{G}}{2}, \frac{1}{4}+\frac{\operatorname{G}}{2}\bigr\rbrack\bigr\rbrack
Properties #

β(7)=61π7/184320\beta(7) = 61\pi^7/184320 — the 6161 is the Euler number ∣E6∣|E_6|

live ↗
InDirichletBeta(7)\mathrm{DirichletBeta}(7)Out61π7184 320\frac{61π^7}{184\,320}

β(−2k)=E2k/2\beta(-2k) = E_{2k}/2 and β(−(2k+1))=0\beta(-(2k+1)) = 0

live ↗
InDirichletBeta([0,−1,−2,−3,−4])\mathrm{DirichletBeta}(\bigl\lbrack0, -1, -2, -3, -4\bigr\rbrack)Out[12,0,−12,0,52]\bigl\lbrack\frac{1}{2}, 0, -\frac{1}{2}, 0, \frac{5}{2}\bigr\rbrack

β(5)=5π5/1536\beta(5) = 5\pi^5/1536

live ↗
InDirichletBeta(5)\mathrm{DirichletBeta}(5)Out5π51 536\frac{5π^5}{1\,536}

Zero at negative odd integers

live ↗
InDirichletBeta(−5)\mathrm{DirichletBeta}(-5)Out00

Known value · Wolfram MathematicalFunctionData, DirichletBeta

β(−4)=52\beta(-4) = \frac52, an Euler number over 2

live ↗
InDirichletBeta(−4)\mathrm{DirichletBeta}(-4)Out52\frac{5}{2}

Known value · Wolfram MathematicalFunctionData, DirichletBeta

β(−3)=0\beta(-3) = 0, a trivial zero

live ↗
InDirichletBeta(−3)\mathrm{DirichletBeta}(-3)Out00

Known value · Wolfram MathematicalFunctionData, DirichletBeta

β(−2)=−12\beta(-2) = -\frac12

live ↗
InDirichletBeta(−2)\mathrm{DirichletBeta}(-2)Out−12-\frac{1}{2}

Known value · Wolfram MathematicalFunctionData, DirichletBeta

The first trivial zero, β(−1)=0\beta(-1) = 0

live ↗
InDirichletBeta(−1)\mathrm{DirichletBeta}(-1)Out00

Known value · Wolfram MathematicalFunctionData, DirichletBeta

β(0)=12\beta(0) = \frac12

live ↗
InDirichletBeta(0)\mathrm{DirichletBeta}(0)Out12\frac{1}{2}

Known value · Wolfram MathematicalFunctionData, DirichletBeta

Possible issues #

Even s≥4s \ge 4 has no closed form and stays symbolic; N() gives 0.98894…0.98894\ldots

live ↗
InDirichletBeta(4)\mathrm{DirichletBeta}(4)OutDirichletBeta(4)\mathrm{DirichletBeta}(4)
About
Examples24
Checked againstWolfram Language 21 agree3 differSymPy 20 agree1 differ1 unevaluatedmpmath 19 agree1 differJulia 19 agree1 error
CrosswalkWikipediaMathWorldWikidataWolfram LanguagempmathJulia

Implementation

referenceenginenotatio
nativeenginetypescriptpackages/ce-patches/src/compute-engine/numerics/dirichlet.ts
mappedexternalwolfram / mpmathpackages/oracle/src/mappings.ts

DirichletBeta[s]; mpmath.dirichlet(s, [0, 1, 0, -1]).

mappedmpmathpackages/oracle/src/mappings.ts

The Dirichlet L-function of the character mod 4.

mappedjuliapackages/oracle/src/mappings.ts
mappedsympypackages/oracle/src/mappings.ts

SymPy has no Dirichlet beta; this is its polylog definition β(s) = (Li_s(i) − Li_s(−i)) / 2i.

See also: DirichletEta, HurwitzZeta, ClausenCl, Zeta