Fact-checked by Grok 2 weeks ago

Legendre function

Legendre functions are a class of special functions in mathematics that arise as solutions to Legendre's differential equation, a second-order linear ordinary differential equation of the form (1 - x^2) y'' - 2x y' + [\nu(\nu + 1) - \mu^2 / (1 - x^2)] y = 0, where \nu is the degree and \mu is the order.^[1] These functions generalize Legendre polynomials P_n(x), which occur when \nu = n (a non-negative integer) and \mu = 0, and associated Legendre functions P_n^m(x), defined for integer m with |m| \leq n.^[2] They also include second-kind solutions like Q_\nu(x) and conical functions, extending to complex arguments and non-integer parameters for broader analytic properties.^[3] Named after the French mathematician Adrien-Marie Legendre (1752–1833), these functions were first introduced in 1783 in his memoir on the attraction of ellipsoids, where he used power series expansions to compute gravitational potentials at exterior points, building on earlier work by Colin Maclaurin.^[4] Legendre's development addressed problems in celestial mechanics and potential theory, earning praise from Pierre-Simon Laplace and contributing to his election as an adjoint of the Paris Academy of Sciences in 1783.^[4] Over time, the functions evolved through contributions from mathematicians like Carl Gustav Jacobi and August Ferdinand Möbius, who generalized them to associated forms in the 19th century. Legendre functions exhibit key properties such as orthogonality over [-1, 1] for polynomials, enabling expansions similar to Fourier series, and recurrence relations that facilitate computation and analysis. For instance, the Legendre polynomials satisfy P_0(x) = 1, P_1(x) = x, and higher degrees via Rodrigues' formula P_n(x) = \frac{1}{2^n n!} \frac{d^n}{dx^n} [(x^2 - 1)^n]. These attributes make them orthogonal basis functions in Hilbert spaces, with generating functions like \frac{1}{\sqrt{1 - 2xt + t^2}} = \sum_{n=0}^\infty P_n(x) t^n for |t| < 1. In applications, Legendre functions are fundamental to solving partial differential equations in spherical coordinates via separation of variables, particularly Laplace's equation for electrostatics and gravitation.^[3] They form the angular components of spherical harmonics Y_l^m(\theta, \phi), where associated Legendre functions P_l^m(\cos \theta) describe the \theta-dependence, essential in quantum mechanics for hydrogen atom wavefunctions and angular momentum operators.^[5] Additional uses span electromagnetism for multipole expansions, geophysical modeling of Earth's gravitational field, and numerical methods in computational physics.^[3]

Mathematical Foundations

Legendre's Differential Equation

The Legendre differential equation is a second-order linear ordinary differential equation that defines the Legendre functions. In its general form, known as the associated Legendre equation, it is given by

(1 - x^2) y'' - 2x y' + \left[ \nu(\nu + 1) - \frac{\mu^2}{1 - x^2} \right] y = 0,

where \nu denotes the degree and \mu the order, both typically complex parameters, though often taken as real or integers in applications.^[1] When \mu = 0, this reduces to the standard Legendre equation

(1 - x^2) y'' - 2x y' + \nu(\nu + 1) y = 0,

which serves as the foundational case for the unassociated Legendre functions.^[1] This equation was introduced by Adrien-Marie Legendre in 1782 as part of his work on the gravitational attraction of ellipsoids, where the functions arose in the series expansion of the Newtonian potential for spheroidal mass distributions.^[4] The equation exhibits regular singular points at x = \pm 1 and at infinity, with indicial exponents \{\pm \frac{1}{2} \mu\} at the finite singularities and \{\nu + 1, -\nu\} at infinity, classifying it as a Fuchsian equation.^[1] The Legendre equation can be transformed into the Gauss hypergeometric differential equation via the substitution t = (1 - x)/2, which maps the interval x \in (-1, 1) to t \in (0, 1) and facilitates expressing solutions in terms of hypergeometric functions. The two linearly independent solutions to the associated equation are conventionally denoted P_\nu^\mu(x) and Q_\nu^\mu(x).^[2]

General Solutions via Hypergeometric Functions

The Legendre differential equation, along with its associated form, possesses regular singular points at x = \pm 1 and at infinity. To derive general analytic solutions, the method of Frobenius is applied around these points, particularly at infinity via the substitution x = 1/z to expand near z = 0. This yields an indicial equation whose roots are \nu + 1 and -\nu, determining the leading behaviors of the solutions as |x| \to \infty. These exponents reflect the Fuchsian nature of the equation and facilitate the construction of series expansions that converge in appropriate domains.^[6] The resulting series solutions can be expressed in terms of the Gauss hypergeometric function {}_2F_1(a, b; c; z). For the associated Legendre function of the first kind, one solution is given by

P_\nu^\mu(x) = \frac{1}{\Gamma(1 - \mu)} \left( \frac{1 + x}{1 - x} \right)^{\mu/2} {}_2F_1\left( -\nu, \nu + 1; 1 - \mu; \frac{1 - x}{2} \right),

up to normalization constants that ensure standard conventions. This representation arises from the transformation t = (1 - x)/2, which maps the interval x \in (-1, 1) to t \in (0, 1) and converts the differential equation into the standard hypergeometric form. The second linearly independent solution is the associated Legendre function of the second kind Q_\nu^\mu(x), which involves a more complex combination of hypergeometric functions to account for the logarithmic singularity at the branch points. The general solution to the associated Legendre equation is thus y(x) = A P_\nu^\mu(x) + B Q_\nu^\mu(x), where A and B are arbitrary constants.^[2] The hypergeometric series {}_2F_1(a, b; c; t) converges absolutely for |t| < 1, corresponding to x > -1 in the transformed variable, with radius of convergence 1 in the t-plane. For non-integer \nu and \mu, the functions exhibit branch points at x = \pm 1, necessitating analytic continuation beyond the principal domain using connection formulas or integral representations to define single-valued branches on the complex plane, often with cuts along [-1, 1]. This continuation preserves linear independence between P_\nu^\mu and Q_\nu^\mu across the Riemann surface. In the specific case \mu = 0, the associated form reduces to the standard Legendre equation, yielding the Legendre function of the first kind P_\nu(x) = {}_2F_1\left( -\nu, \nu + 1; 1; \frac{1 - x}{2} \right), with the second solution Q_\nu(x) ensuring completeness. For integer \nu = n, this specializes to polynomial solutions, but the hypergeometric form remains valid for general \nu.

Legendre Functions of the First Kind

Definition and Explicit Forms for Integer Order

The Legendre functions of the first kind for integer order, denoted P_n(x) with n \geq 0 an integer, are polynomial solutions of degree n to Legendre's differential equation with azimuthal order \mu = 0, uniquely determined by the normalization condition P_n(1) = 1.^[7] These polynomials form an orthogonal basis on the interval [-1, 1] with respect to the weight function w(x) = 1, and their leading coefficient is \frac{2^n (\frac{1}{2})_n}{n!}, where (\frac{1}{2})_n denotes the Pochhammer symbol.^[7] One explicit representation is given by Rodrigues' formula:

P_n(x) = \frac{1}{2^n n!} \frac{d^n}{dx^n} (x^2 - 1)^n.

^[8] This formula generates the polynomials directly from repeated differentiation of the binomial (x^2 - 1)^n, ensuring the required degree and normalization. An equivalent form uses (1 - x^2)^n with an adjusted sign factor, but the version above is conventional for deriving further properties.^[9] The generating function for the sequence \{P_n(x)\}_{n=0}^\infty is

\sum_{n=0}^\infty P_n(x) t^n = \frac{1}{\sqrt{1 - 2xt + t^2}},

valid for |t| < 1 and x \in [-1, 1].^[10] This closed-form expression facilitates the expansion of functions in terms of Legendre polynomials and highlights their role in potential theory and spherical harmonics. For low orders, the polynomials take simple forms: P_0(x) = 1, P_1(x) = x, and P_2(x) = \frac{1}{2}(3x^2 - 1).^[9] These examples illustrate the progression from constants to higher-degree terms with increasing powers of x. The Legendre polynomials exhibit definite parity: P_n(-x) = (-1)^n P_n(x), making them even functions for even n and odd for odd n.^[9] Additionally, they are bounded on the interval of orthogonality, satisfying |P_n(x)| \leq 1 for all x \in [-1, 1], with equality at the endpoints due to the normalization.^[8]

General Order and Branch Cuts

The Legendre function of the first kind for general real or complex order \nu is defined via analytic continuation of the hypergeometric representation, extending beyond integer degrees where it reduces to polynomials. For -1 < x < 1, it is given by

P_\nu(x) = {}_2F_1\left(-\nu, \nu+1; 1; \frac{1-x}{2}\right),

where {}_2F_1 denotes the Gauss hypergeometric function.^[2] This expression provides the principal branch, which is real-valued when \nu and x are real. As \nu approaches a non-negative integer n, P_\nu(x) limits to the Legendre polynomial P_n(x). The domain of the principal branch is the interval (-1, 1), with analytic continuation to the complex plane excluding branch cuts along the real axis from -\infty to 1. This cut structure arises from the singularities at the endpoints x = \pm 1 and at infinity, ensuring single-valuedness within the principal sheet. For |x| > 1, values are obtained via connection formulas that account for the monodromy around the branch points.^[11] For associated Legendre functions, the Ferrers function of the first kind \mathbf{P}_\nu^\mu(x) serves as a variant suited to the interval -1 < x < 1, defined as

\mathbf{P}_\nu^\mu(x) = \left(\frac{1+x}{1-x}\right)^{\mu/2} \mathbf{F}\left(\nu+1, -\nu; 1-\mu; \frac{1}{2} - \frac{x}{2}\right),

where \mathbf{F} is the regularized hypergeometric function and \mu, \nu \in \mathbb{R}. This form, often expressed in terms of P_\nu^{-\mu}(x) in some conventions, incorporates a prefactor that ensures real values for real arguments in the principal domain and facilitates applications in toroidal coordinates.^[2] The Ferrers function relates to the standard associated Legendre function P_\nu^\mu(x) by a transformation involving (1 - x^2)^{\mu/2}, highlighting normalization differences: Legendre functions emphasize polynomial behavior for integer orders, while Ferrers variants prioritize boundedness and reality on (-1, 1).^[2] For large |\nu|, asymptotic approximations of P_\nu(\cos \theta) are obtained using the Mehler-Dirichlet integral representation:

P_\nu(\cos \theta) = \frac{\sqrt{2}}{\pi} \int_0^\theta \frac{\cos\left( \left(\nu + \frac{1}{2}\right) t \right) \, dt}{\sqrt{\cos t - \cos \theta}},

valid for $0 < \theta < \pi. This integral allows evaluation via methods such as stationary phase, yielding leading-order behavior proportional to \nu^{-1/2} near \theta = 0 and oscillatory decay elsewhere, which is crucial for high-frequency wave problems.^[12] Uniqueness of these functions for general \nu holds up to the choice of normalization, with Legendre and Ferrers conventions differing primarily in phase factors and prefactors for associated cases.^[2]

Legendre Functions of the Second Kind

Definition and Explicit Forms

The Legendre functions of the second kind, denoted Q_\nu(x), provide the second linearly independent solution to Legendre's differential equation, complementing the functions of the first kind P_\nu(x). For x > 1, an explicit representation is given by

Q_\nu(x) = \frac{1}{2} P_\nu(x) \ln \left( \frac{x+1}{x-1} \right) - W_{\nu-1}(x),

where W_{\nu-1}(x) is a series expressible in terms of hypergeometric functions or other special functions, ensuring the form captures the singular behavior.^[13] For integer orders n = 0, 1, 2, \dots, the expression simplifies, with W_{n-1}(x) becoming a polynomial of degree n-1:

Q_n(x) = \frac{1}{2} P_n(x) \ln \left( \frac{x+1}{x-1} \right) - \sum_{k=1}^n \frac{1}{k} P_{n-k}(x) P_{k-1}(x).

This sum arises from the connection formulas and ensures Q_n(x) is well-defined for computation.^[14] These functions are analytic in the domain |x| > 1, where the principal branch of the logarithm is taken, but exhibit logarithmic singularities at the branch points x = \pm 1, reflecting their role in solutions with singular behavior.^[15] A key normalization occurs for n = 0:

Q_0(x) = \frac{1}{2} \ln \left( \frac{x+1}{x-1} \right),

which follows directly from the general formula with P_0(x) = 1 and W_{-1}(x) = 0.^[16] The functions Q_\nu(x) and P_\nu(x) are linearly independent over the complex plane, as their Wronskian is nonzero, forming a fundamental pair for solving Legendre's equation.

Behavior at Singular Points

The Legendre functions of the second kind, Q_\nu(x), exhibit singular behavior at the regular singular points x = \pm 1 of Legendre's differential equation, characterized by logarithmic divergences that distinguish them from the functions of the first kind. Near x = 1^-, for fixed \nu \not\in \{-1, -2, \dots\}, the leading asymptotic behavior is Q_\nu(x) \sim -\frac{1}{2} \ln(1 - x), with higher-order terms involving constants such as Euler's constant \gamma and the digamma function \psi(\nu + 1). This logarithmic singularity arises from the integral representation or hypergeometric series expansion of Q_\nu(x), ensuring the function is unbounded as the argument approaches the endpoint of the interval (-1, 1). A similar logarithmic divergence occurs near x = -1^+, where the coefficient depends on \nu through connection formulas relating values across the branch cut, typically yielding Q_\nu(x) \sim \frac{1}{2} P_\nu(-1) \ln(1 + x) + subleading terms, with P_\nu(-1) = \cos(\pi \nu) providing the parity adjustment for integer orders. This behavior reflects the symmetric placement of singular points in the differential equation, with the precise coefficient incorporating \nu to maintain consistency with recurrence relations. At infinity, for large |x| with x > 1, Q_\nu(x) decays algebraically as Q_\nu(x) \sim \frac{\sqrt{\pi}}{\Gamma(\nu + 3/2) (2x)^{\nu+1}}, providing the dominant term in the expansion that ensures convergence in applications requiring solutions at large distances. This asymptotic form highlights the polynomial-like decay modulated by \nu. The singular behaviors at x = \pm 1 are linked to the monodromy of solutions around these branch points, as described by the Riemann P-symbol for Legendre's equation, which encodes the local exponents $0, 0 at each finite singular point and the branching structure for non-integer \nu.^[17] This monodromy analysis reveals how encircling the points \pm 1 induces logarithmic phase shifts in Q_\nu(x), contrasting with the single-valued nature of P_\nu(x) near these points. In comparison, the Legendre functions of the first kind P_\nu(x) remain finite at x = \pm 1 for all \nu, with P_\nu(1) = 1 and P_\nu(-1) = \cos(\pi \nu), thus completing an independent basis of solutions without singularities in the physical domain [-1, 1] for integer orders.

Associated Legendre Functions

Definitions for First and Second Kind

The associated Legendre functions generalize the Legendre functions of the first and second kinds by introducing a non-zero order parameter \mu, which arises naturally in problems with azimuthal dependence, such as those in spherical coordinates. These functions satisfy the associated Legendre differential equation,

(1 - x^2) \frac{d^2 w}{dx^2} - 2x \frac{d w}{dx} + \left[ \nu(\nu + 1) - \frac{\mu^2}{1 - x^2} \right] w = 0,

where \nu is the degree and \mu is the order, typically taken as integers in many applications.^[1] For integer \mu \geq 0, the associated Legendre function of the first kind is defined as

P_\nu^\mu(x) = (-1)^\mu (1 - x^2)^{\mu/2} \frac{d^\mu}{dx^\mu} P_\nu(x),

where P_\nu(x) is the Legendre function of the first kind. This definition ensures the function is regular at x = \pm 1 for appropriate \nu and \mu, and it reduces to the standard Legendre function when \mu = 0. For integer degree n and order m with |m| \leq n, P_n^m(x) becomes a polynomial of degree n - m, often referred to as an associated Legendre polynomial. A representative example is P_1^1(x) = -(1 - x^2)^{1/2}.^[18]^[19] The associated Legendre function of the second kind, for integer \mu \geq 0, is similarly defined as

Q_\nu^\mu(x) = (-1)^\mu (1 - x^2)^{\mu/2} \frac{d^\mu}{dx^\mu} Q_\nu(x),

where Q_\nu(x) is the Legendre function of the second kind. This function exhibits logarithmic singularities at x = \pm 1 and is used to form the complete set of solutions to the associated Legendre equation. An illustrative example is Q_0^1(x) = (1 - x^2)^{1/2} / (x^2 - 1).^[13] In spherical coordinates, the associated Legendre functions of the first kind are integral to the definition of spherical harmonics, where Y_l^m(\theta, \phi) \propto P_l^{|m|}(\cos \theta) e^{i m \phi}, providing the \theta-dependent part of the angular solutions to Laplace's equation in three dimensions.^[20]

Normalization and Symmetry Properties

The normalization of associated Legendre functions of the first kind, P_\nu^\mu(x), often incorporates the Condon-Shortley phase in physics contexts to ensure orthonormality when combined with azimuthal factors in spherical harmonics. Specifically, for integer orders n and m \geq 0, the convention is P_n^m(x) = (-1)^m (1 - x^2)^{m/2} \frac{d^m}{dx^m} P_n(x), where P_n(x) is the Legendre polynomial of degree n. This phase factor (-1)^m simplifies matrix elements in angular momentum calculations and aligns with the unit normalization of spherical harmonics over the unit sphere, \int Y_l^{m*}(\theta, \phi) Y_l^{m'}(\theta, \phi) \, d\Omega = \delta_{ll'} \delta_{mm'}. Symmetry properties of these functions include parity relations that facilitate their use in even-odd decompositions. For nonnegative integers n and m with m \leq n, P_n^m(-x) = (-1)^{n+m} P_n^m(x). The relation to negative order is given by P_\nu^{-\mu}(x) = (-1)^\mu \frac{\Gamma(\nu - \mu + 1)}{\Gamma(\nu + \mu + 1)} P_\nu^\mu(x), ensuring consistency across positive and negative \mu. Alternative normalizations appear in specialized fields, such as geophysics, where Ferrers functions P_n^m(x) are defined without the Condon-Shortley phase as P_n^m(x) = (1 - x^2)^{m/2} \frac{d^m}{dx^m} P_n(x) for |x| < 1, often paired with Schmidt quasi-normalization for surface spherical harmonics to handle real-valued geomagnetic data efficiently.^[21] For imaginary orders \mu = i \tau (\tau real), Ferrers functions extend to toroidal coordinates, aiding in geophysical modeling of Earth's magnetic field. A key reflection formula connects functions of degree \nu and -\nu - 1: P_\nu^\mu(x) = P_{-\nu-1}^\mu(x), which is essential for analytic continuation and reciprocity in potential theory.^[22] In quantum applications, half-integer orders arise in contexts like the Dirac equation for relativistic particles, where functions such as P_{l + 1/2}^\mu(x) exhibit behaviors tied to hypergeometric representations, maintaining orthogonality under specific boundary conditions while preserving the overall symmetry structure.

Representations and Expansions

Series Expansions

The Legendre functions of the first kind P_\nu(x) possess a series expansion expressible through the Gauss hypergeometric function, facilitating numerical computation within appropriate domains. Specifically,

P_\nu(x) = \, _2F_1\left(-\nu, \nu+1; 1; \frac{1-x}{2}\right),

where the hypergeometric function expands as the power series

_2F_1(a,b;c;z) = \sum_{k=0}^\infty \frac{(a)_k (b)_k}{(c)_k k!} z^k,

with Pochhammer symbols (a)_k = a(a+1)\cdots(a+k-1) and (a)_0 = 1. Substituting the parameters yields

P_\nu(x) = \sum_{k=0}^\infty \frac{(-\nu)_k (\nu+1)_k}{k!^2} \left( \frac{1-x}{2} \right)^k.

This series converges for \left| \frac{1-x}{2} \right| < 1, equivalent to \operatorname{Re}(x) > 0, and extends analytically to the complex plane excluding the branch cut (-\infty, 1]. For integer orders \nu = n \geq 0, the series terminates at k = n due to the vanishing Pochhammer symbol (-\nu)_k = 0 for k > n, reducing P_n(x) to a polynomial of degree n. For the Legendre functions of the second kind Q_\nu(x), the expansion incorporates a logarithmic singularity reflecting the branch point at x = \pm 1. For x > 1,

Q_\nu(x) = \frac{\sqrt{\pi} \, \Gamma(\nu+1)}{2^{\nu+1} \Gamma\left(\nu + \frac{3}{2}\right)} x^{-\nu-1} \, _2F_1\left( \frac{\nu+1}{2}, \frac{\nu+2}{2}; \nu + \frac{3}{2}; \frac{1}{x^2} \right),

which provides a power series in $1/x^2 converging for x > 1. For integer orders n \geq 0 and x > 1,

Q_n(x) = \frac{1}{2} P_n(x) \ln \left( \frac{x+1}{x-1} \right) - \sum_{k=1}^n \frac{1}{k} P_{k-1}(x) P_{n-k}(x),

where the finite sum involves polynomials of lower degree, enabling efficient evaluation. This form arises from the limiting case of the general hypergeometric representation and highlights the non-polynomial nature of Q_n(x), diverging logarithmically as x \to 1^+. For -1 < x < 1, the principal value uses the Ferrers function of the second kind, which includes similar logarithmic and hypergeometric components but requires careful branch handling. Associated Legendre functions of the first kind P_\nu^\mu(x) also admit hypergeometric series expansions. For integer orders n and m with |m| \leq n,

P_n^m(x) = (-1)^m \left(1 - x^2\right)^{m/2} \frac{(n+m)!}{2^m m! (n-m)!} \, _2F_1\left(-n+m, n+m+1; m+1; \frac{1-x}{2}\right),

converging under the same condition \operatorname{Re}(x) > 0. For general \nu, \mu, the form generalizes accordingly, with termination occurring when parameters cause Pochhammer symbols to vanish. Similar expansions hold for associated functions of the second kind Q_\nu^\mu(x), often combining hypergeometric terms with logarithmic factors.

Integral Representations

Integral representations of Legendre functions trace their origins to the late 18th-century developments in potential theory by Adrien-Marie Legendre and Pierre-Simon Laplace, who employed such forms to expand the gravitational potential of spheroids and ellipsoids in series and integrals for celestial mechanics problems. These early Laplace integrals facilitated the solution of Poisson's equation in spherical coordinates, laying the groundwork for modern analytic expressions that enable evaluation and continuation beyond integer orders. A prominent real integral representation for the Legendre function of the first kind, known as the Mehler-Dirichlet integral, is given by

P_\nu(\cos \theta) = \frac{\sqrt{2}}{\pi} \int_0^\theta \frac{\cos\left((\nu + \frac{1}{2})\phi\right)}{\sqrt{\cos \phi - \cos \theta}} \, d\phi,

valid for $0 < \theta < \pi and complex \nu with appropriate convergence conditions. This formula, derived by Dirichlet in 1836 and generalized by Mehler in 1847, proves especially effective for analytic continuation to non-integer \nu and for deriving asymptotic behaviors in the complex plane. For general complex arguments, the Schläfli contour integral offers a powerful representation for P_\nu(z):

P_\nu(z) = \frac{1}{2\pi i} \oint \frac{(t^2 - 1)^\nu}{2^\nu (t - z)^{\nu + 1}} \, dt,

where the closed contour encircles the branch cut [-1, 1] in the positive sense, assuming z outside the cut and suitable branch choices for the multi-valued functions. Introduced by Ludwig Schläfli in 1850, this integral circumvents singularities and supports analytic continuation across branch cuts, making it ideal for computational purposes in regions away from the real interval [-1, 1]. The Legendre function of the second kind Q_\nu(z) possesses a real-line integral representation linking it directly to P_\nu:

Q_\nu(z) = \frac{1}{2} \int_{-1}^1 \frac{P_\nu(t)}{z - t} \, dt,

for z \notin [-1, 1], interpreted as a Cauchy principal value when z approaches the cut. This form, arising from the theory of singular integrals and orthogonal expansions, facilitates the computation of Q_\nu via known values of P_\nu and is instrumental in asymptotic expansions for large |z|. For associated Legendre functions, the Schläfli integral generalizes to

P_\nu^\mu(z) = \frac{\Gamma(\nu - \mu + 1)}{\Gamma(\nu + 1)} \frac{1}{2^\nu \pi i} (z^2 - 1)^{-\mu/2} \oint \frac{(t^2 - 1)^\nu}{(t - z)^{\nu + \mu + 1}} \, dt,

with the contour encircling [-1, 1] and branches chosen consistently; a similar form holds for negative order P_\nu^{-\mu}(z). This representation extends the utility of contour integrals to associated cases, aiding in the study of toroidal and conical functions in potential problems. These integral forms are particularly valuable for deriving asymptotic approximations in high-order or large-argument regimes.

Key Properties

Recurrence Relations and Differentiation Formulas

Legendre functions satisfy a variety of recurrence relations that enable the computation of higher-order functions from lower-order ones, facilitating numerical evaluation and analytical manipulations. For the Legendre polynomials P_n(x) of integer degree n, the fundamental three-term recurrence relation is

(n+1) P_{n+1}(x) = (2n+1) x P_n(x) - n P_{n-1}(x),

with initial conditions P_0(x) = 1 and P_1(x) = x. This relation, derived from the hypergeometric representation or the differential equation, allows recursive generation of the polynomials and holds for |x| \leq 1. Differentiation formulas for P_n(x) express the derivative in terms of the polynomials themselves. One such relation is

\frac{d}{dx} P_n(x) = \frac{n}{x^2 - 1} \left( x P_n(x) - P_{n-1}(x) \right),

valid for x^2 \neq 1, which follows from combining the recurrence with the Legendre differential equation. An equivalent form is (1 - x^2) P_n'(x) = n P_{n-1}(x) - n x P_n(x). For Legendre functions of general (non-integer) degree \nu, a key relation involving derivatives is

\nu P_\nu(x) = x P_\nu'(x) - P_{\nu-1}'(x).

This identity arises from logarithmic differentiation of the hypergeometric series expansion and is useful in asymptotic analysis and solutions to boundary value problems.^[23] The associated Legendre functions P_\nu^\mu(x) obey similar recurrences adjusted for the order \mu. A prominent differentiation formula is

(1 - x^2) \frac{d}{dx} P_\nu^\mu(x) = \nu x P_\nu^\mu(x) - (\nu + \mu) P_{\nu-1}^\mu(x),

which generalizes the integer-degree case and supports derivations in spherical coordinate systems. This can be paired with the recurrence (\nu - \mu + 2) P_{\nu+2}^\mu(x) - (2\nu + 3) x P_{\nu+1}^\mu(x) + (\nu + \mu + 1) P_\nu^\mu(x) = 0 for computational stability. Sums of Legendre functions are connected via the Christoffel-Darboux identity, a summation formula for orthogonal polynomials. For Legendre polynomials, it states

\sum_{k=0}^n \frac{2k + 1}{2} P_k(x) P_k(y) = \frac{n+1}{2} \frac{P_{n+1}(x) P_n(y) - P_n(x) P_{n+1}(y)}{x - y},

for x \neq y, with the confluent form at x = y involving derivatives: \sum_{k=0}^n \frac{2k + 1}{2} P_k(x)^2 = \frac{n+1}{2} (P_{n+1}'(x) P_n(x) - P_n'(x) P_{n+1}(x)). This identity, fundamental for kernel representations and quadrature, extends to associated functions with appropriate weights.

Orthogonality and Completeness for Integer Orders

The Legendre polynomials P_n(x) for nonnegative integers n form an orthogonal set on the interval [-1, 1] with respect to the constant weight function 1. Specifically, their orthogonality relation is given by

\int_{-1}^{1} P_m(x) P_n(x) \, dx = \frac{2}{2n + 1} \delta_{mn},

where \delta_{mn} is the Kronecker delta, equal to 1 if m = n and 0 otherwise.^[7] This property extends to the associated Legendre functions P_l^m(x) of integer order, where l \geq |m| \geq 0 are integers. For fixed m, the functions P_l^m(x) are orthogonal on [-1, 1] with

\int_{-1}^{1} P_l^m(x) P_k^m(x) \, dx = \frac{2}{2l + 1} \frac{(l + m)!}{(l - m)!} \delta_{lk}.

The Legendre polynomials constitute a complete orthogonal basis for the Hilbert space L^2[-1, 1] equipped with the inner product \langle f, g \rangle = \int_{-1}^{1} f(x) g(x) \, dx. Consequently, any function f \in L^2[-1, 1] admits a Fourier-Legendre series expansion

f(x) = \sum_{n=0}^{\infty} a_n P_n(x),

where the coefficients are a_n = \int_{-1}^{1} f(x) P_n(x) \, dx, and the series converges to f in the L^2 norm.^[24] Associated with this completeness is Parseval's identity, which equates the L^2 norm of f to a sum over the squared coefficients:

\int_{-1}^{1} |f(x)|^2 \, dx = \sum_{n=0}^{\infty} \frac{2n+1}{2} |a_n|^2.[](https://dlmf.nist.gov/18.2) In contrast, the Legendre functions of the second kind $Q_\nu^\mu(x)$ exhibit logarithmic singularities at the endpoints $x = \pm 1$, precluding standard orthogonality relations on $[-1, 1]$ in the $L^2$ sense.[](https://dlmf.nist.gov/14.8) ## Applications ### Potential Theory and Multipole Expansions In potential theory, Legendre functions play a fundamental role in describing axisymmetric solutions to [Laplace's equation](/page/Laplace's_equation), particularly for problems involving gravitational or electrostatic potentials. Adrien-Marie [Legendre](/page/Adrien-Marie_Legendre) introduced these functions in his 1782 memoir on the gravitational attraction of ellipsoids, where he developed methods to compute the potential due to homogeneous ellipsoidal mass distributions at exterior points, laying the groundwork for their use in [celestial mechanics](/page/celestial_mechanics) and [geophysics](/page/geophysics). A key application arises in the expansion of the reciprocal distance between two points, which represents the fundamental solution to Laplace's equation in three dimensions. For points $\mathbf{r}$ and $\mathbf{r}'$ with magnitudes $r$ and $r'$, and angle $\gamma$ between them, the expansion is given by \[ \frac{1}{|\mathbf{r} - \mathbf{r}'|} = \sum_{n=0}^{\infty} \frac{r_<^n}{r_>^{n+1}} P_n(\cos \gamma),

where r_< = \min(r, r') and r_> = \max(r, r'), and P_n are the Legendre polynomials.^[25] This series converges for r \neq r' and enables the decomposition of potentials from distributed sources into separable angular and radial components./04:_Series_Solutions/4.05:_Legendre_Polynomials) For axisymmetric boundary value problems, solutions to Laplace's equation \nabla^2 \Phi = 0 in spherical coordinates (r, \theta, \phi) take the form \Phi(r, \theta) = \sum_{n=0}^{\infty} \left( A_n r^n + \frac{B_n}{r^{n+1}} \right) P_n(\cos \theta), assuming azimuthal independence.^[26] The terms with positive powers of r describe interior solutions (regular at the origin), while the inverse powers handle exterior solutions (vanishing at infinity).^[27] Coefficients A_n and B_n are determined by boundary conditions, such as specified potentials on spheres or other surfaces of revolution. In electrostatics, this framework underpins the multipole expansion of the potential due to a localized charge distribution. The leading term for n=0 is the monopole, proportional to the total charge Q and scaling as $1/r, representing the net Coulomb potential.^[28] The n=1 dipole term involves the dipole moment \mathbf{p} and falls as $1/r^2, capturing the potential asymmetry for neutral distributions with separated charges.^[29] Higher-order terms, such as the n=2 quadrupole, describe further deviations and decay as $1/r^3, essential for precise modeling of molecular or planetary fields. Legendre series also facilitate numerical solutions to boundary value problems in potential theory, where the potential on a boundary is expanded as \Phi(\theta) = \sum_{n=0}^{\infty} C_n P_n(\cos \theta), and coefficients are found via orthogonality integrals to approximate the full solution inside or outside the domain. This approach is particularly effective for axisymmetric geometries, reducing the problem to solving for radial coefficients after angular decomposition. For non-axisymmetric cases, associated Legendre functions extend this method to include azimuthal dependence.

Quantum Mechanics and Spherical Harmonics

In quantum mechanics, associated Legendre functions play a central role in describing the angular dependence of wave functions for systems with spherical symmetry, particularly through their incorporation into spherical harmonics. These functions arise as solutions to the angular part of the Schrödinger equation in spherical coordinates, where the separation of variables leads to the associated Legendre equation for the polar angle θ. The spherical harmonics Y_l^m(\theta, \phi), which form a complete orthonormal basis for functions on the sphere, are defined as

Y_l^m(\theta, \phi) = \sqrt{ \frac{(2l+1)(l - m)!}{4\pi (l + m)!} } \, P_l^m(\cos \theta) \, e^{i m \phi},

where P_l^m are the associated Legendre functions, l is the orbital angular momentum quantum number (l = 0, 1, 2, \dots), and m is the magnetic quantum number (-l \leq m \leq l). This normalization ensures the orthogonality relation \int Y_l^{m*}(\theta, \phi) Y_{l'}^{m'}(\theta, \phi) \, d\Omega = \delta_{ll'} \delta_{mm'}, with the integral over the solid angle d\Omega = \sin\theta \, d\theta \, d\phi.^[30] A prime example is the hydrogen atom, where the total wave function \psi_{nlm}(r, \theta, \phi) = R_{nl}(r) Y_l^m(\theta, \phi) separates into radial and angular parts. The radial function R_{nl}(r) involves associated Laguerre polynomials, while the angular part relies on the associated Legendre functions via the spherical harmonics to capture the quantum states' dependence on direction. This structure reflects the conservation of angular momentum, with L^2 Y_l^m = \hbar^2 l(l+1) Y_l^m and L_z Y_l^m = \hbar m Y_l^m, enabling the classification of atomic orbitals (e.g., s, p, d) by l and m. The associated Legendre functions ensure the wave functions vanish appropriately at the poles (\theta = 0, \pi) for m \neq 0, maintaining physical boundary conditions.^[31]^[32] The addition theorem for spherical harmonics connects Legendre polynomials to multipole expansions in quantum contexts, such as scattering or perturbation theory:

P_l(\cos \gamma) = \frac{4\pi}{2l + 1} \sum_{m=-l}^l Y_l^{m*}(\theta', \phi') Y_l^m(\theta, \phi),

where \cos \gamma is the angle between directions (\theta, \phi) and (\theta', \phi'). This relation facilitates the decomposition of rotationally invariant operators and is crucial for calculating matrix elements in angular momentum coupling. The orthogonality of spherical harmonics directly implies selection rules for quantum transitions, such as electric dipole radiation, where \Delta l = \pm 1 and \Delta m = 0, \pm 1 arise from the integral \int Y_{l'}^{m'*} \, \mathbf{r} \, Y_l^m \, d\Omega vanishing unless these conditions hold, governing allowed spectral lines in atomic spectra.^[5]^[33]^[34]

Advanced Topics

Legendre Functions as Characters in Harmonic Analysis

In the representation theory of the Lie group SL(2,ℝ), Legendre functions arise as zonal spherical functions on the symmetric space SL(2,ℝ)/SO(2), which is isometric to the hyperbolic plane ℍ². Specifically, for the principal series representations parameterized by ν ∈ ℝ, the spherical function φ_ν(g) is given by P_{-1/2 + iν}(cosh t), where t is the hyperbolic distance from the identity coset e·SO(2) to g·SO(2), and P_μ denotes the Legendre function of the first kind. These functions satisfy the spherical function equation derived from the Casimir operator of the group algebra, confirming their role as the Legendre equation solutions in this context.^[35] As characters, P_ν(cosh t) traces the irreducible unitary representations induced from the Borel subgroup, providing the radial part of the matrix coefficients in the K-biinvariant functions on the group.^[36] In Fourier analysis on SL(2,ℝ)/SO(2), the Legendre transform serves as the Harish-Chandra transform for radial functions, decomposing them into spherical harmonics via integration against P_ν(cosh t). The Plancherel formula for this decomposition states that for a K-biinvariant function f on SL(2,ℝ),

\|f\|_{L^2}^2 = \int_{-\infty}^{\infty} |\hat{f}(\nu)|^2 \frac{\nu \tanh(\pi \nu / 2)}{4} \, d\nu + \text{discrete terms},

where \hat{f}(ν) is the Legendre transform \hat{f}(ν) = \int_0^{\infty} f(t) P_{-1/2 + iν}(cosh t) \sinh t , dt, and the measure involves |P_ν|^2 implicitly through the inversion.^[37] This formula arises from the decomposition of L²(SL(2,ℝ)/SO(2)) into principal series representations, with the continuous spectrum parameterized by ν.^[37] The connection to spherical functions on general symmetric spaces follows Harish-Chandra's integral formula, which for rank-one spaces like SL(2,ℝ)/SO(2) specializes to the explicit form involving Legendre functions. Harish-Chandra's c-function c(λ) = Γ(1/2 + iλ) Γ(1/2 - iλ) / [2π Γ(iλ) Γ(-iλ)] appears in the inversion, yielding the density |c(λ)|^{-2} in the Plancherel measure.^[38] This framework extends Legendre functions to higher-rank semisimple groups, where they generalize to Jacobi functions, but retain the Legendre form for the SL(2,ℝ) case.^[38]

Singularities and Symmetry Considerations

The Legendre functions of the first kind, P_\nu(x), for non-integer degree \nu are bounded and real-valued on the interval [-1, 1], though they exhibit oscillatory behavior and lack the even/odd parity of integer-degree polynomials. This arises from the branch cuts in their analytic continuation outside [-1, 1], stemming from the hypergeometric representation, with cuts typically along (-\infty, -1] and [1, \infty). Contrasting with the bounded polynomials for integer \nu, these functions are defined via converging series on the closed interval.^[2] The mirror symmetry of Legendre's differential equation, which remains invariant under the transformation x \to -x, imposes significant constraints on the solutions. For integer \nu, this symmetry results in even or odd polynomials that respect the equation's reflection principle, ensuring boundedness. However, for non-integer \nu, the solutions do not preserve strict parity, leading to asymmetric behavior. This symmetry consequence manifests in the complex plane through specific pole structures, notably at orders \nu = -1/2 + i \tau for real \tau > 0, where the functions develop features tied to the equation's regular singular points at x = \pm 1.^[39] The monodromy group of the Legendre equation, generated by analytic continuations around the branch points at x = \pm 1 and infinity, is finite for integer \nu, reflecting the polynomial nature of the solutions. For non-integer \nu, the multi-valuedness introduces a more complex monodromy group, connected to the broader framework of differential Galois theory, where the Galois group of the Picard-Vessiot extension captures the algebraic structure of the solutions and their symmetries. This distinction underscores the algebraic rigidity for integer cases versus the transcendental complexity otherwise.^[40] A prominent example of such non-integer order functions is the conical (or Mehler) function P_{-1/2 + i \tau}(x), which arises in applications involving axial symmetry, such as toroidal coordinates. These functions feature infinite oscillatory behavior along the branch cuts from -\infty to -1 and $1 to \infty in the complex x-plane, resulting from the imaginary component in the order parameter. For -1 < x < 1, they remain real and suitable for numerical computation.^[41] In physical applications, such as potential theory and quantum mechanics, non-integer orders are generally avoided in favor of integer \nu to ensure polynomial solutions with orthogonality properties on [-1, 1], preserving physical interpretability. This selection criterion aligns with the symmetry requirements of the underlying problems, where polynomial Legendre functions provide the necessary regularity.^[39]

References

[1]
DLMF: §14.2 Differential Equations ‣ Real Arguments ‣ Chapter 14 Legendre and Related Functions
### Definition of Legendre Functions from §14.2
[2]
DLMF: §14.3 Definitions and Hypergeometric Representations ...
Ferrers functions, associated Legendre functions, definitions, hypergeometric function, hypergeometric representations, relations to other functions ...
[3]
DLMF: Chapter 14 Legendre and Related Functions
Chapter 14 Legendre and Related Functions. T. M. Dunster Department of Mathematics and Statistics, San Diego State University, San Diego, California.
[4]
Adrien-Marie Legendre - Biography - MacTutor
He then introduced what we call today the Legendre functions and used these to determine, using power series, the attraction of an ellipsoid at any exterior ...
[5]
[PDF] The Spherical Harmonics
which relates the Legendre polynomials to the spherical harmonics with m = 0. In terms of the spherical harmonics, the general solution to Laplace's ...
[6]
[PDF] Legendre's Polynomials Chapter-4
1 transfers the singular point x = 1 to t = 0. This transformed differential equation is in the hypergeometric form with a = − n , b = n + 1 and c = 1.
[7]
DLMF: §18.3 Definitions ‣ Classical Orthogonal Polynomials ...
Legendre polynomials are special cases of Legendre functions, Ferrers functions, and associated Legendre functions (§14.7(i)). In consequence, additional ...Missing: mathematics | Show results with:mathematics
[8]
https://arxiv.org/pdf/2210.10942.pdf
[9]
DLMF: §18.5 Explicit Representations ‣ Classical Orthogonal ...
For corresponding formulas for Chebyshev, Legendre, and the Hermite 𝐻𝑒 n polynomials apply (18.7. 3)–(18.7. 6), (18.7.<|control11|><|separator|>
[10]
DLMF: §18.9 Recurrence Relations and Derivatives ‣ Classical Orthogonal Polynomials ‣ Chapter 18 Orthogonal Polynomials
### Summary of Generating Function for Legendre Polynomials
[11]
DLMF: §14.24 Analytic Continuation ‣ Complex Arguments ...
14 Legendre and Related FunctionsComplex Arguments14.23 Values on the Cut14.25 Integral Representations ... branch point 1 (but not the branch point − 1 ) ...
[12]
DLMF: §14.12 Integral Representations ‣ Real Arguments ...
Mehler–Dirichlet Formula ; Γ · ( z ) : gamma function ; 𝖯 ν μ · ( x ) : Ferrers function of the first kind ; cos · z : cosine function ; x : differential of x ; ℜ · : ...
[13]
DLMF: §14.7 Integer Degree and Order ‣ Real Arguments ‣ Chapter 14 Legendre and Related Functions
### Summary of Legendre Functions of the Second Kind $ Q_\nu(x) $ from §14.7
[14]
https://dlmf.nist.gov/14.7.E4
[15]
https://dlmf.nist.gov/14.8
[16]
Legendre functions of fractional degree: transformations and ...
Apr 1, 2016 · Legendre's differential equation (2.1) on the Riemann sphere P z 1 is of the 'hypergeometric' sort, with only three singular points, z=±1 ...
[17]
Associated Legendre Functions - Richard Fitzpatrick
Associated Legendre Functions. The associated Legendre functions, $ P_l^{\,m}(x)$ , are the well-behaved solutions of the differential equation ...
[18]
DLMF: §14.6 Integer Order ‣ Real Arguments ‣ Chapter 14 Legendre and Related Functions
### Summary of Associated Legendre Functions of First Kind for Integer m
[19]
DLMF: §14.30 Spherical and Spheroidal Harmonics ‣ Applications ‣ Chapter 14 Legendre and Related Functions
### Summary: Relation Between Associated Legendre Functions and Spherical Harmonics
[20]
Geomagnetism and Schmidt quasi-normalization - Oxford Academic
Ferrers normalized functions are therefore unsuitable for numerical work, producing coefficients an,m and bn,m, which have a wide range of numerical values ...
[21]
Legendre Polynomial -- from Wolfram MathWorld
... Legendre polynomials, but defined on the interval (0, 1). They obey the ... int_(-1)^1xP_L(x)P_N(x)dx, = {(2(L+1))/((2L+1). (55). int_(-1)^1x^2P_L(x)P_N(x) ...Missing: ≤ | Show results with:≤
[22]
DLMF: §14.17 Integrals ‣ Real Arguments ‣ Chapter 14 Legendre ...
Orthogonality relations for the associated Legendre functions of imaginary order are given in Bielski (2013).Missing: Qnu( | Show results with:Qnu(
[23]
DLMF: §18.18 Sums ‣ Classical Orthogonal Polynomials ‣ Chapter ...
... Rodrigues formula (Table 18.5.1), integration by parts, and (18.5.12). (18.18.13) follows from (18.18.12) for α = ± 1 2 by (18.7.19), (18.7.20). Permalink ...
[24]
18.2 General Orthogonal Polynomials
If the polynomials p n ⁡ ( x ) ( n = 0 , 1 , … , N ) are orthogonal on a finite set X of N + 1 distinct points as in (18.2.3), then the polynomial p N + 1 ⁡ ...
[25]
DLMF: §14.8 Behavior at Singularities ‣ Real Arguments ‣ Chapter 14 Legendre and Related Functions
### Summary of Applications of Legendre Functions (DLMF §14.8)
[26]
[PDF] Expansion of 1/r potential in Legendre polynomials - Physics
Page 1. Expansion of 1/r potential in Legendre polynomials. • In electrostatics and gravitation, we see scalar potentials of the form V = K d. • Take d = |R −r| ...
[27]
[PDF] 154 - 4.6. Solutions of Laplace's Equation in Spherical
A general solution of Laplace's equation in spherical polar coordinates, assum- ing axial symmetry, is therefore the following: ŽA„r¹P„(cos 00 v = Σ А„r²P₂(cos ...
[28]
[PDF] LAPLACE'S EQUATION IN SPHERICAL COORDINATES
We know that the solutions to the Legendre equation are the Legendre polynomials, Pl (cos θ).
[29]
[PDF] Multipole Expansion of the Electrostatic Potential - UT Physics
... Legendre polynomials. Moreover, the Legendre polynomials are normalized so that P`(1) = 1 for all. `, hence. F(Θab = θa) = 2` + 1. 4π. × P`(cosθa). (70). This ...Missing: theory | Show results with:theory
[30]
[PDF] Multipole Expansion of the Electrostatic Potential - UT Physics
Multipole Expansion of the Electrostatic Potential. Mathematical Background. Let me start with a bit of mathematical theorem: Consider two points with ...
[31]
[PDF] Legendre functions, Spherical Harmonics, and Bessel Functions
Legendre polynomials and legendre functions more generally solve the θ equations. Bessel functions arise in problems with spherical symmetry, but actually ...
[32]
[PDF] 4.5 The Hydrogen Atom
The spherical harmonics with negative azimuthal number -m can be expressed in terms of those with positive azimuthal number m. Y -m. 1. (ϑ,ϕ)=(−1) m.
[33]
[PDF] Chapter 10 The Hydrogen Atom The Schrodinger Equation in ...
We choose to include a factor of (−1)m with the associated Legendre polynomials, and the sign of all spherical harmonics will be positive as a result.
[34]
18. Electrostatics Using Spherical Coordinates: Spherical Harmonics
... ∇2 acting on the surface of a sphere. Addition Theorem. An important identity is the so-called addition theorem for spherical harmonics: Pl(cosγ)=4π2l+1l ...
[35]
[PDF] Quantum Mechanics II - University of Tennessee, Knoxville
Since spherical harmonics are orthogonal this expression can be simplified. ... The equations (91) (92) are the selection rules for the elecric dipole transition.
[36]
[PDF] arXiv:2401.04271v4 [quant-ph] 11 Jun 2024
Jun 11, 2024 · A key component is the CNOT gate that preserves the rank of spherical tensor operators. ... of quantum computing [13–18]. The conventional approaches for FTQC are ...
[37]
[PDF] arXiv:2211.15872v1 [quant-ph] 29 Nov 2022
Nov 30, 2022 · basis of spherical tensor operators ordered by rank. We presented a numerical analysis of two paradigmatic models of quantum chaos in both the many-body and.
[38]
[PDF] Differential equations and the algebra of confluent spherical ...
Oct 1, 2015 · We define q ∈ Q(gC) to be spherical whenever q = ˜q. The above system of differential equations have been extensively used by Harish-Chandra in ...
[39]
[PDF] Representations of SL2(R) - UBC Math
Nov 1, 2020 · For G = SL2(R), every irreducible unitary representation of G restricts to a direct sum of characters of. SO2, each occurring at most once. We ...
[40]
[PDF] Harish-Chandra's Plancherel formula for SL(2,R)
Sep 27, 2019 · Let R = R(gC,hC) be the root system for h. Let W(G, H) = NG(H)/ZG(H) be the analytic Weyl group of (G, H).
[41]
[PDF] On Squares of Eigenfunctions for the Hyperbolic Plane and a New ...
With the exception of the odd-dimensional real hyperbolic spaces, all other rank-one spaces have Harish-Chandra transform that involves fractional ...
[42]
[PDF] Mellin transforms of Whittaker functions - Numdam
Let G be a connected reductive algebraic group defined and quasi-split over R. In this paper the Mellin transform of the Whittaker function.
[43]
[PDF] arXiv:2304.09319v3 [math-ph] 20 Oct 2023
Oct 20, 2023 · The formula can also be generalized to the k largest or, similarly, smallest eigenvalues of any random ma- trices when the n-point correlation ...
[44]
The Singularity of Legendre Functions of the First Kind as a ... - MDPI
We present and explain a stand-alone and in-depth argument for rejecting all solutions of Legendre's equation in physics apart from the polynomial ones.Missing: integer | Show results with:integer
[45]
[PDF] hypergeometric equation and monodromy groups
in Riemann's P-symbol: (4.1) w(z) = P... ξ η ζ α β γ ; z α0 β0 γ0 ... The equation is called Legendre's differential equation if m = 0. Its ...
[46]
DLMF: §14.20 Conical (or Mehler) Functions ‣ Real Arguments ...
Solutions are known as conical or Mehler functions. For − 1 < x < 1 and τ > 0 , a numerically satisfactory pair of real conical functions is 𝖯 − 1 2 + i ⁢ τ ...Missing: Dirichlet nu