Hurwitz zeta function

1 Definition and Basic Properties

1.1 Series definition and convergence domain

The Hurwitz zeta function, denoted \(\zeta(s,a)\), is defined for complex \(s\) and a complex parameter \(a\) (commonly taken with \(\Re(a)>0\)) by the Dirichlet series \[ \zeta(s,a)=\sum_{n=0}^{\infty}(n+a)^{-s}. \] For \(\Re(s)>1\), the terms behave like \(n^{-\Re(s)}\), so the series converges absolutely by comparison with the \(p\)-series. In that region it defines a holomorphic function of \(s\) and is smooth in \(a\) away from singular parameter choices.

1.2 Parameter dependence: role of \(a\)

The parameter \(a\) shifts the lattice of summation from \(1,2,3,\dots\) to \(a,a+1,a+2,\dots\). As \(a\) varies, the analytic structure in the \(a\)-plane reflects the positions where terms \((n+a)^{-s}\) hit poles. For many standard properties (such as series manipulations and integral representations), it is convenient to assume \(\Re(a)>0\) so that the defining series and certain transforms can be justified directly, with later analytic continuation extending the statements.

1.3 Simple evaluations and limiting cases

A basic limiting behavior connects the Hurwitz zeta function to the Riemann zeta function. When \(a=1\), \[ \zeta(s,1)=\sum_{n=0}^{\infty}(n+1)^{-s}=\zeta(s), \]

1.4 Analytic continuation and meromorphic structure

Although the series defines \(\zeta(s,a)\) only for \(\Re(s)>1\), the function admits analytic continuation to a meromorphic function of \(s\) on the whole complex plane. Standard constructions rely on integral representations and subtraction/renormalization that match the original series for \(\Re(s)>1\). The resulting meromorphic structure is characterized by a single pole at \(s=1\) under typical parameter assumptions.

1.5 Poles, residues, and special values

For fixed \(a\) with \(\Re(a)>0\), \(\zeta(s,a)\) has a simple pole at \(s=1\). Its residue is independent of \(a\) and equals \(1\). At other integer points, the function takes finite values and is closely related to Bernoulli polynomials and other classical special functions, enabling explicit computations at both positive and nonpositive integers.

2 Analytic Representations

2.1 Integral representations

2.1.1 Mellin transform viewpoint

One common approach views \((n+a)^{-s}\) as the Mellin transform of an exponential. For \(\Re(s)>0\) and suitable \(a\), the identity \[ (n+a)^{-s}=\frac{1}{\Gamma(s)}\int_{0}^{\infty} t^{s-1}e^{-(n+a)t}\,dt \] allows termwise summation over \(n\), producing an integral representation of \(\zeta(s,a)\). After summing the geometric series \(\sum_{n=0}^\infty e^{-nt} = \frac{1}{1-e^{-t}}\), one obtains integrals involving \(\frac{e^{-at}}{1-e^{-t}}\). This form is especially useful for analytic continuation, because divergences near \(t=0\) can be isolated systematically.

2.1.2 Contour and Hankel-type integrals

Beyond the real-line integral, contour methods yield Hankel-type representations that incorporate analytic continuation through branch cuts of \(t^{s-1}\). These contour integrals can be arranged to respect the principal branch choices for complex powers, and they often lead to efficient derivations of functional relations and asymptotics. Such representations are typical in treatments connecting zeta functions with the Gamma function and Mellin inversion.

2.2 Asymptotic expansions

2.2.1 Euler–Maclaurin summation connection

Asymptotic behavior in the parameter \(a\) can be derived from the Euler–Maclaurin summation formula applied to the defining series. The procedure expresses the difference between the sum \(\sum_{n\ge 0} f(n)\) and the corresponding integral \(\int f\) in terms of derivatives and Bernoulli numbers/polynomials. For \(\zeta(s,a)\), this yields expansions of the form \[ \zeta(s,a)\sim \frac{1}{s-1}a^{1-s}+\frac{1}{2}a^{-s}+\sum_{k=1}^{N} \frac{B_{2k}}{(2k)!}\,(s)_{2k-1}\,a^{-(s+2k-1)}, \] where \((s)_{m}\) denotes a rising or falling factorial depending on convention, and truncation errors decrease as \(N\) increases. These expansions explain how \(\zeta(s,a)\) interpolates between power-law leading terms and more rapidly decaying corrections.

2.3 Functional-type relations and shifts

2.3.1 Relation between \(\zeta(s,a)\) and \(\zeta(s,a+1)\)

A basic shift identity follows directly from the definition by separating the first term: \[ \zeta(s,a)=a^{-s}+\sum_{n=1}^\infty(n+a)^{-s}=a^{-s}+\zeta(s,a+1). \] Equivalently, \[ \zeta(s,a+1)=\zeta(s,a)-a^{-s}. \] Such relations connect values at neighboring parameter points and are repeatedly used in proofs and in deriving properties at integer arguments.

2.3.2 Recurrence via derivatives with respect to \(a\)

Differentiation with respect to \(a\) introduces factors of \((n+a)^{-s-1}\). For \(\Re(s)>0\) where termwise differentiation is justified, one finds \[ \frac{\partial}{\partial a}\zeta(s,a)=-s\,\zeta(s+1,a). \] Higher derivatives generate higher shifts in \(s\), giving a recurrence that links parameter calculus to evaluations at shifted \(s\)-values. These formulas underpin connections to polygamma functions and to analytic expansions in \(a\).

3 Special Values and Connections

3.1 Values at positive integers and Bernoulli polynomials

At positive integers, \(\zeta(s,a)\) can be expressed in terms of Bernoulli polynomials when one considers expansions and special limits. While \(\zeta(s,a)\) at \(s=1\) involves a pole, values at integers \(s=m\ge 2\) are finite and correspond to generalized sums that admit closed forms in special cases (e.g., rational \(a\)). The systematic appearance of Bernoulli polynomials is most explicit in formulas for negative integer arguments and in generating functions derived from integral representations.

3.2 Values and expansions at nonpositive integers

3.2.1 Relation to Bernoulli polynomials

A hallmark property states that for nonpositive integers, \(\zeta(-n,a)\) is a polynomial in \(a\) of degree \(n+1\). Specifically, \[ \zeta(-n,a)=-\frac{B_{n+1}(a)}{n+1}\quad (n\ge 0), \] where \(B_{k}(a)\) denotes the Bernoulli polynomials. This relation turns the Hurwitz zeta function into a concise encoding device for Bernoulli polynomials, with the analytic continuation ensuring that these values arise from the meromorphic continuation across the pole at \(s=1\).

3.3 Relation to the Riemann zeta function

3.3.1 The case \(a=1\)

As noted earlier, taking \(a=1\) recovers the classical Riemann zeta function: \[ \zeta(s,1)=\zeta(s). \] This identification places many known features of \(\zeta(s)\)—such as its analytic continuation and its pole structure—inside the broader framework of \(\zeta(s,a)\).

3.3.2 Variants with rational parameters

When \(a\) is rational, \(\zeta(s,a)\) can often be expressed via linear combinations of Riemann zeta values and related Dirichlet \(L\)-functions. The general mechanism uses periodicity when splitting the summation into arithmetic progressions. Such decompositions are important for explicit computation and for connecting zeta values to arithmetic objects, while also providing practical formulas for numerical evaluation.

3.4 Connection to the gamma function and logarithmic derivatives

3.4.1 Links involving \(\log\Gamma(a)\)

The Hurwitz zeta function ties to the Gamma function through limits and derivatives that produce logarithmic Gamma terms. A standard example appears when considering the behavior near \(s=0\) or derivatives with respect to \(s\), where \(\log\Gamma(a)\) emerges naturally from the same integral representations that define the zeta function. This link is central in regularization contexts (for instance, converting divergent products into finite quantities) even when such applications are treated only at an overview level here.

3.4.2 Derivatives at selected points

Derivatives of \(\zeta(s,a)\) with respect to \(s\) at special points (notably \(s=0\) and negative integers) yield expressions involving \(\log\Gamma(a)\) and generalized logarithmic constants. These derivatives also generate series expansions useful for both theoretical analysis and computation, because they isolate the “finite part” after analytic continuation.

4 Derivatives, Expansions, and Parameter Calculus

4.1 Differentiation with respect to \(s\)

For \(\Re(s)>1\), differentiation under the sum gives \[ \frac{\partial}{\partial s}\zeta(s,a)=-\sum_{n=0}^\infty (n+a)^{-s}\log(n+a), \] and higher derivatives involve higher powers of \(\log(n+a)\). Through analytic continuation, these derivatives extend to meromorphic or holomorphic functions at broader parameter ranges (with careful treatment at pole locations such as \(s=1\)). Derivative values at integer points are key inputs for expansions and special-function identities.

4.2 Differentiation with respect to \(a\)

As described in the shift section, differentiation with respect to \(a\) produces shifted zeta values: \[ \frac{\partial}{\partial a}\zeta(s,a)=-s\,\zeta(s+1,a). \] This identity gives a direct “raising operator” in \(s\) and makes it straightforward to build expansions in \(a\), especially when combined with Taylor series around points where \(\zeta(s,a)\) is regular.

4.3 Taylor and Laurent expansions around special points

The meromorphic nature in \(s\) implies that around the pole at \(s=1\), \(\zeta(s,a)\) admits a Laurent expansion with a principal part \(\frac{1}{s-1}\) and a finite remainder. Around regular points, it has Taylor expansions. In applications, one often isolates constant terms and coefficients of \((s-1)^k\), because these coefficients can be related to generalized constants and Bernoulli polynomial data.

4.4 Stieltjes-type constants for the Hurwitz zeta function

Analogous to the Stieltjes constants for the Riemann zeta function, the Hurwitz zeta function has a family of constants defined from the expansion near \(s=1\): \[ \zeta(s,a)=\frac{1}{s-1}+\sum_{n=0}^{\infty}\frac{(-1)^n}{n!}\gamma_n(a)\,(s-1)^n. \] The coefficients \(\gamma_n(a)\) generalize the classical Stieltjes constants and depend on \(a\). They appear naturally in analytic number theory and in computations of finite parts of zeta regularizations.

5 Computational Aspects

5.1 Numerical evaluation strategies

5.2 Acceleration methods and series acceleration

When direct summation converges slowly, convergence acceleration techniques can be employed. These include Euler–Maclaurin-based truncations (using Bernoulli polynomial correction terms) and techniques that reorganize the series into more rapidly decaying components. In practice, one balances the cost of additional correction terms against the reduced number of required summation steps.

5.3 Truncation error estimates for common representations

Any numerical method needs reliable control of truncation error. For series truncation, tail bounds follow from comparison with integrals of \(x^{-\Re(s)}\) or from known remainder estimates in Euler–Maclaurin formulas. For integral representations, one estimates the contributions from the small-\(t\) and large-\(t\) portions separately, ensuring that the chosen contour or real-line path yields controllable error.

5.4 Handling complex arguments and branch choices

When \(s\) or \(a\) is complex, computations must specify branches for multi-valued functions such as \((n+a)^{-s}=\exp(-s\log(n+a))\). Consistent branch selection for \(\log(n+a)\) is crucial to avoid discontinuities. Robust algorithms also track phase conventions in Gamma functions and any contour-based integrals used for continuation.

6 Further Theory and Applications

6.1 Relation to polygamma functions

The Hurwitz zeta function connects closely to the polygamma functions through differentiation identities. In particular, the logarithmic derivative structure of \(\Gamma(a)\) and its higher derivatives can be expressed using \(\zeta(s,a)\) at shifted arguments. These relations allow one to translate properties between the zeta function and the better-studied family of polygamma functions, including recurrence and asymptotic behavior.

6.2 Spectral and regularization applications (overview-level)

In spectral theory, zeta functions arise from sums over eigenvalues or from determinants defined through analytic continuation. The Hurwitz zeta function is a common building block when eigenvalue sequences resemble arithmetic progressions or when boundary conditions shift the spectrum by a parameter \(a\). In regularization, one uses analytic continuation to assign finite values to divergent series, and the residue/finite-part structure near poles provides the constants needed to define renormalized quantities.

6.3 Modular or transformation behavior (general perspective)

While \(\zeta(s,a)\) is not itself a modular form in the usual sense, it participates in transformation patterns related to shifts, reflections, and functional equations derived from theta-function methods or contour transforms. These general “functional-type” behaviors reflect the analytic continuation and symmetry properties of the underlying integral representations.

6.4 Occurrences in power series and generating functions

Hurwitz zeta values appear in expansions of special functions and in generating-function contexts. Because Bernoulli polynomials provide closed forms at nonpositive integers, one can often convert zeta-generated coefficients into polynomial or series data. Additionally, parameter derivatives produce logarithmic terms that enter generating functions for constants and for asymptotic coefficients in related analytic problems.