General integral-identities involving Legendre polynomials and their derivatives
Yannick Wunderlich, Kyungseon Joo, Victor I. Mokeev · 2025
Read General integral-identities involving Legendre polynomials and their derivatives by Yannick Wunderlich, Kyungseon Joo, Victor I. Mokeev (2025) 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-2509-22904.txt.
Introduction
Yannick Wunderlich, Kyungseon Joo, and Victor I Mokeev work out integrals that mix Legendre polynomials with their derivatives. Such integrals show up in multipole expansions and in spectral methods for physics. Yet closed forms for any derivative order are hard to find in books or software. The note fills that gap.
The three authors use repeated integration by parts and parity arguments to reach explicit formulas. Their results come as binomial and Gamma-function forms that hold for all valid index choices. The formulas merge and extend known orthogonality rules. They aim to give ready tools for both hand work and code.
Catalog record
| CREATOR | Yannick Wunderlich, Kyungseon Joo, Victor I. Mokeev |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 34 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-22904
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
General integral-identities involving Legendre polynomials and their derivatives Yannick Wunderlich, Kyungseon Joo, Victor I. Mokeev Abstract Integrals involving derivatives of Legendre polynomials frequently arise in applications ranging from multipole expansions for processes involving electromagnetic probes (cf. BallPhotoproduction ; KDT ) to spectral methods in numerical physics. Despite their practical relevance, closed-form expressions for such integrals - particularly involving arbitrary derivative orders - are not readily accessible in standard references or symbolic tools. In this note, we derive and present general analytic expressions for integrals of the form ∫ − 1 + 1 d x P n ( q ) ( x ) P m ( k ) ( x ) , \int_{-1}^{+1}dxP^{(q)}_{n}(x)P^{(k)}_{m}(x), where P n ( x ) P_{n}(x) and P m ( x ) P_{m}(x) are Legendre polynomials and q q , k k denote their order of differentiation. Using repeated integration by parts, parity arguments, and closed-form boundary evaluations, we obtain explicit binomial and Gamma-function representations valid for all non-negative integers n n , m m , q q , k k . These results unify and extend known orthogonality relations and provide ready-to-use tools for analytic and computational contexts. I Introduction Integrals involving Legendre polynomials and their derivatives arise in a broad range of physical and mathematical contexts - from multipole expansions in classical electrodynamics, quantum angular momentum theory and partial-wave analysis (see e.g. formalisms in BallPhotoproduction ; KDT ), to modern applications in spectral methods and orthogonal polynomial techniques for numerical solutions of differential equations. Despite their prevalence in such domains, closed-form expressions for integrals involving derivatives of Legendre polynomials - especially at arbitrary differentiation orders - are surprisingly absent from standard references. Classical resources such as Abramowitz & Stegun AbramowitzStegun , the Digital Library of Mathematical Functions dlmf , and comprehensive integral tables like Gradshteyn & Ryzhik gradryzhik or Prudnikov, Brychkov & Marichev prudnikov typically provide orthogonality relations and product integrals involving the polynomials themselves, but offer only limited results for derivative overlaps. The work QureshiEtAl has considered special cases, particularly involving first- and second-order derivatives, but these tend to focus on either symmetric cases (e.g., \int P_{n}^{\prime}(x)P_{n}^{\prime}(x)\,dx ) or specific recursion-based identities. More recent treatments, such as kholshevnikov , explore integrals of iterated Legendre functions or related generating functions, but still do not provide a general framework for integrals of the form \int_{-1}^{+1}dxP^{(q)}_{n}(x)P^{(k)}_{m}(x), (1) where q , k are arbitrary non-negative integers denoting derivative orders. This note aims to provide a unified and accessible derivation of such integrals (for all non-negative integers n,m,q,k ) by employing a combination of integration by parts, parity arguments, boundary evaluations, and factorial/Gamma-function identities. As a key ingredient, we present a general formula for the boundary values P_{n}^{(k)}(x)\big|_{x=+1} , namely P_{n}^{(k)}(x)\big|_{x=+1}=\frac{(n+k)!}{2^{k}\,k!\,(n-k)!}, (2) which we derive using only an elementary recurrence relation and a known uniqueness-argument from discrete mathematics (cf. Appendix A of this note). While the resulting expression may be known in some form, we are not aware of any equally direct and pedagogically motivated derivation in the standard literature. Our goal is not to claim originality of the final results obtained in this work, but to present them in a form that is compact, verifiable, and useful for both analytical and numerical applications. II Overlap of ordinary Legendre polynomial with 1st order derivative We begin by proving the following result for the integral of a Legendre polynomial times the derivative of another Legendre polynomial: \int_{-1}^{+1}dxP_{n}(x)P^{\prime}_{m}(x)=\begin{cases}0,&n+m\text{ even,}\\ 0,&n+m\text{ odd and }m<n,\\ 2,&n+m\text{ odd and }n<m.\end{cases} (3) Proof: Via integration by parts, we can show relatively quickly that: \int_{-1}^{+1}dxP_{n}(x)P^{\prime}_{m}(x)=\left[P_{n}(x)P_{m}(x)\right]\big|_{-1}^{+1}-\int_{-1}^{+1}dxP^{\prime}_{n}(x)P_{m}(x). (4) Now, at the boundaries x=\pm 1 , it is well-known that Legendre polynomials satisfy (due to definite parity of the polynomials, as well as normalization at x=+1 ): P_{n}(+1)=1\text{, and }P_{n}(-1)=(-1)^{n}. (5) Thus, our integral becomes \int_{-1}^{+1}dxP_{n}(x)P^{\prime}_{m}(x)=\left[1-(-1)^{m+n}\right]-\int_{-1}^{+1}dxP^{\prime}_{n}(x)P_{m}(x). (6) For m+n even, we have \left[1-(-1)^{m+n}\right]=0 and therefore \int_{-1}^{+1}dxP_{n}(x)P^{\prime}_{m}(x)=-\int_{-1}^{+1}dxP^{\prime}_{n}(x)P_{m}(x). (7) Now to show that for this case, our considered integral is indeed zero, we have to consider further parity-arguments. Start with P_{n}(-x)=(-1)^{n}P_{n}(x). (8) For the derivative, we obtain the following due to the chain rule: P^{\prime}_{n}(-x)=\frac{d}{d(-x)}P_{n}(-x)=\frac{dx}{d(-x)}\frac{d}{dx}(-1)^{n}P_{n}(x)=(-1)^{n+1}P^{\prime}_{n}(x). (9) Thus, for our integrand f(x):=P_{n}(x)P^{\prime}_{m}(x) , the parity relation becomes: f(-x)=P_{n}(-x)P^{\prime}_{m}(-x)=(-1)^{n}P_{n}(x)(-1)^{m+1}P^{\prime}_{m}(x)=(-1)^{m+n+1}f(x), (10) and we see directly that for even values of m+n , our integrand is odd and thus the integral vanishes. Now, let us return to our general relation ( 4 ) and assume that on the left-hand-side there, we have m+n odd and n<m satisfied. Thus, we get \int_{-1}^{+1}dxP_{n}(x)P^{\prime}_{m}(x)=\left[1-(-1)^{m+n}\right]-\underbrace{\int_{-1}^{+1}dxP^{\prime}_{n}(x)P_{m}(x)}_{m+n\text{ odd and }n<m\text{ here}}. (11) Since the Legendre polynomials form a complete orthogonal system of functions, we know that for the derivative P^{\prime}_{n}(x) under the integral in the right-hand-side, the following re-expansion formula holds: P^{\prime}_{n}(x)=\sum_{j=0}^{n-1}a^{(n-1)}_{j}P_{j}(x), (12) with some as of yet unspecified coefficients a^{(n-1)}_{j} . We know that a general expansion of the form ( 12 ) has to exist and terminate at the order \mathcal{O}(x^{n-1}) , since the derivative P^{\prime}_{n}(x) is a polynomial of the same order. However, the re-expansion ( 12 ) into Legendre-polynomials without derivatives involves only ordinary Legendre polynomials P_{j}(x) of order j<n . Thus, we also have j<m for all j , and the integral appearing on the right-hand-side of equation ( 11 ) vanishes, due to basic and well-known orthogonality properties of Legendre polynomials. We furthermore get \underbrace{\int_{-1}^{+1}dxP_{n}(x)P^{\prime}_{m}(x)}_{{}_{m+n\text{ odd and }n<m\text{ here}}}=\left[1-(-1)^{m+n}\right]=1-(-1)=2, (13) i.e. the integral is in this case fully fixed by the boundary term ’ \left[1-(-1)^{m+n}\right] ’. This completes our derivation of the general result ( 3 ). QED. The result ( 3 ) can also be written in a more concise form using the Heaviside \theta -function: \int_{-1}^{+1}dxP_{n}(x)P^{\prime}_{m}(x)=\theta(m-n)\left[1-(-1)^{m+n}\right]. (14) We notice that keeping the bracket ’ \left[1-(-1)^{m+n}\right] ’ ensures a ’parity-filter’. We have to declare here that within this note, we use the so-called ’right-continuous’ Heaviside-function, i.e. \theta(x)\equiv\Theta_{>}(x)=\begin{cases}1,&x>0,\\ 0,&x\leq 0.\end{cases}, (15) such that boundary-cases are excluded. III Overlap of ordinary Legendre polynomial with 2nd order derivative In many cases, one would also need the integral involving the second derivative of a Legendre-polynomial: \int_{-1}^{+1}dxP_{n}(x)P^{\prime\prime}_{m}(x). (16) Here, one can again do integration by parts, but now two successive times: \displaystyle\int_{-1}^{+1}dxP_{n}(x)P^{\prime\prime}_{m}(x) \displaystyle=\left[P_{n}(x)P^{\prime}_{m}(x)\right]\big|_{-1}^{+1}-\int_{-1}^{+1}dxP^{\prime}_{n}(x)P^{\prime}_{m}(x) \displaystyle=\left[P_{n}(x)P^{\prime}_{m}(x)\right]\big|_{-1}^{+1}-\left[P^{\prime}_{n}(x)P_{m}(x)\right]\big|_{-1}^{+1}+\int_{-1}^{+1}dxP^{\prime\prime}_{n}(x)P_{m}(x). (17) For the evaluation of the boundary term, one has to know that Legendre’s differential equation yields (this is a special case of the more generally formula ( 35 ), which is re-derived in Appendices A and B ) P^{\prime}_{n}(x=+1)=\frac{n(n+1)}{2}, (18) and therefore one also gets P^{\prime}_{n}(x=-1)=(-1)^{n+1}P^{\prime}_{n}(x=+1)=(-1)^{n+1}\frac{n(n+1)}{2}. (19) With all this, we get for the first summand in the boundary-term \left[P_{n}(x)P^{\prime}_{m}(x)\right]\big|_{-1}^{+1}=1\cdot\frac{m(m+1)}{2}-(-1)^{n}\cdot(-1)^{m+1}\frac{m(m+1)}{2}=\frac{m(m+1)}{2}\left[1-(-1)^{n+m+1}\right], (20) and for the second one similarly \left[P^{\prime}_{n}(x)P_{m}(x)\right]\big|_{-1}^{+1}=\frac{n(n+1)}{2}\cdot 1-(-1)^{n+1}\frac{n(n+1)}{2}\cdot(-1)^{m}=\frac{n(n+1)}{2}\left[1-(-1)^{n+m+1}\right], (21) We thus effectively consider the integral: \int_{-1}^{+1}dxP_{n}(x)P^{\prime\prime}_{m}(x)=\left(\frac{m(m+1)}{2}-\frac{n(n+1)}{2}\right)\left[1-(-1)^{n+m+1}\right]+\int_{-1}^{+1}dxP^{\prime\prime}_{n}(x)P_{m}(x). (22) We now repeat the parity-argument from above, but with a new integrand \tilde{f}(x):=P_{n}(x)P^{\prime\prime}_{m}(x) , which satisfies the parity-relation: \tilde{f}(-x)=P_{n}(-x)P^{\prime\prime}_{m}(-x)=(-1)^{n}P_{n}(x)\cdot(-1)^{m+2}P^{\prime\prime}_{m}(x)=(-1)^{n+m}\tilde{f}(x). (23) This directly tells us that our considered integral vanishes for odd values of m+n . Furthermore, we automatically get the result \underbrace{\int_{-1}^{+1}dxP_{n}(x)P^{\prime\prime}_{m}(x)}_{{}_{m+n\text{ even and }n\geq(m-1)\text{ here}}}=0, (24) due to ’re-expansion arguments’ for the Legendre-polynomial P^{\prime\prime}_{m}(x) under the integral (similar to arguments given below equation ( 11 ) in section II ). Therefore, in our further considerations, we consider the case where n<(m-1) and again can do this kind of ’trick’ where the boundary-term effectively fixes the value of the integral, since the remaining integral on the right-hand-side vanishes: \displaystyle\underbrace{\int_{-1}^{+1}dxP_{n}(x)P^{\prime\prime}_{m}(x)}_{m+n\text{