1 Formulation of Legendre’s differential equation

1.1 Standard differential equation form

Legendre’s differential equation is the second-order linear ordinary differential equation \[ (1-x^2)\,y''-2x\,y'+\ell(\ell+1)\,y=0, \] where \(y=y(x)\) and \(x\) is the independent variable. The parameter \(\ell\) is commonly taken to be real, and in physical applications it often corresponds to an angular-momentum index.

A central feature of the equation is its regular singular structure at \(x=\pm 1\), reflecting the geometry of spherical coordinates.

1.2 Normalization and parameter conventions

Different texts use slightly different conventions. One common convention is that \(\ell\) is an integer \(n\ge 0\) when polynomial solutions are sought. Under this choice, the equation admits solutions that are polynomials of degree \(n\). Another convention permits \(\ell\) to be non-integer; in that case, solutions are generally not polynomials and typically involve special functions.

Normalization also varies by author. The standard normalization for Legendre polynomials is \[ P_n(1)=1, \] which uniquely fixes the scaling among polynomial solutions.

1.3 Relationship to Sturm–Liouville theory

Legendre’s differential equation can be written in Sturm–Liouville form: \[ \frac{d}{dx}\!\left[(1-x^2)\frac{dy}{dx}\right]+\ell(\ell+1)\,y=0. \] This identification places the problem within the classical framework of orthogonal eigenfunctions. With appropriate boundary conditions at \(x=\pm 1\), different admissible values of \(\ell\) yield an orthogonal set of eigenfunctions with respect to the weight \(w(x)=1\) on \([-1,1]\).

2 Solution methods

2.1 Power-series (Frobenius) approach

A standard technique is to expand the solution as a power series (or Frobenius series) about a point such as \(x=0\), which is an ordinary point. One assumes a series \[ y(x)=\sum_{k=0}^{\infty} a_k x^k. \] Substituting into the differential equation yields a recursion relation among coefficients \(a_k\). The resulting series can be analyzed for termination (producing a polynomial) or for general non-terminating behavior.

2.1.1 Recurrence relations for series coefficients

Carrying out the substitution produces a three-term relation linking \(a_{k+2}\) to earlier coefficients. The structure enforces a parity separation: coefficients of even and odd powers decouple. Concretely, the recursion implies that solutions can be classified as even or odd functions depending on the initial choice of \(a_0\) or \(a_1\).

2.1.2 Conditions for polynomial solutions

For the series to terminate at some finite degree \(n\), the parameter \(\ell\) must satisfy \(\ell=n\) with \(n\in\mathbb{N}_0\). Under this condition, the recurrence forces higher-degree coefficients to vanish, yielding a polynomial solution \(P_n(x)\) (up to an overall constant factor). If \(\ell\) is not a nonnegative integer, termination generally does not occur, and the solutions remain non-polynomial.

2.2 Rodrigues’ formula connection

Polynomial solutions can also be derived using Rodrigues’ formula: \[ P_n(x)=\frac{1}{2^n n!}\frac{d^n}{dx^n}\left[(x^2-1)^n\right]. \] This expression provides a direct construction of the polynomial sequence and makes it straightforward to compute derivatives and integrals involving \(P_n\). It also clarifies the parity and endpoint behavior.

2.3 Generating functions

Legendre polynomials admit a generating function: \[ \frac{1}{\sqrt{1-2xt+t^2}}=\sum_{n=0}^{\infty} P_n(x)\,t^n, \]

valid for \(t<1\) (for real \(x\)). Expanding the right-hand side and comparing coefficients yields explicit formulas and recurrence relations. Generating functions are also useful for summing series and studying asymptotic behavior.

2.4 Hypergeometric-function representation

More general solutions can be expressed with hypergeometric functions. In particular, one may write a polynomial solution for integer \(\ell=n\) in terms of a terminating hypergeometric series. For general \(\ell\), hypergeometric representations provide analytic continuation beyond the interval \([-1,1]\) and describe how solutions behave near singular points.

3 Legendre polynomials

3.1 Polynomial solutions and degrees

When \(\ell=n\) is a nonnegative integer, the solution space contains a unique (up to scaling) polynomial of degree \(n\) that is regular on \([-1,1]\). This polynomial is the Legendre polynomial \(P_n(x)\). Its degree and normalization are fixed by the choice \(P_n(1)=1\).

3.2 Low-degree examples and patterns

The first few polynomials are: \[ P_0(x)=1,\quad P_1(x)=x,\quad P_2(x)=\frac{1}{2}(3x^2-1),\quad P_3(x)=\frac{1}{2}(5x^3-3x), \] and so on. A recurring pattern is evident: \(P_n\) has parity matching \(n\) (even \(n\) gives an even polynomial; odd \(n\) gives an odd polynomial), and the leading term grows as \(x^n\) with a coefficient ensuring the normalization at \(x=1\).

3.3 Orthogonality on \([-1,1]\)

Legendre polynomials form an orthogonal family on \([-1,1]\) with respect to the weight \(w(x)=1\). Specifically, \[ \int_{-1}^{1} P_m(x)\,P_n(x)\,dx = 0 \quad \text{for } m\ne n. \] This orthogonality is a direct consequence of the Sturm–Liouville structure and underpins projection formulas for expanding functions into Legendre series.

3.4 Weight function and inner product

With weight \(w(x)=1\), the associated inner product is \[ \langle f,g\rangle=\int_{-1}^{1} f(x)\,g(x)\,dx. \] Under this inner product, the set \(\{P_n\}_{n\ge 0}\) becomes orthogonal, and the norm is \[ \int_{-1}^{1} [P_n(x)]^2\,dx=\frac{2}{2n+1}. \] These relations define a natural basis for representing sufficiently regular functions on \([-1,1]\).

3.5 Normalization conventions

While \(P_n(1)=1\) is standard, other normalizations appear in applied literature, sometimes choosing monic polynomials or scaling by \(\sqrt{(2n+1)/2}\) to make the basis orthonormal. Translating between conventions typically involves multiplying by a constant factor depending on \(n\).

4 Second independent solution (non-polynomial)

4.1 General behavior of the second solution

For a fixed \(\ell\), the differential equation has two linearly independent solutions. When \(\ell=n\) is a nonnegative integer, one solution is the polynomial \(P_n(x)\). The second independent solution is non-polynomial and is generally singular at one or both endpoints \(x=\pm 1\). It is often denoted \(Q_n(x)\), up to normalization.

4.2 Regular vs singular solutions on \([-1,1]\)

Because \(x=\pm 1\) are regular singular points, one solution is typically regular (finite) at both endpoints only in the polynomial case, while the other tends to diverge (logarithmically for integer orders). This distinction is crucial in boundary-value problems, where physical or geometric constraints select the regular branch.

4.3 Physical and boundary-condition selection

In applications involving spherical symmetry, common boundary requirements (boundedness, finite energy, or smoothness across a domain) eliminate the non-polynomial solution. As a result, the eigenfunctions used in expansions are often exactly the Legendre polynomials rather than the singular second solution.

5 Recurrence relations and identities

5.1 Three-term recurrence relation

Legendre polynomials satisfy a three-term recurrence: \[ (n+1)P_{n+1}(x)=(2n+1)x\,P_n(x)-n\,P_{n-1}(x). \] This relation enables efficient generation of polynomials without resorting to repeated differentiation, and it is a cornerstone for computational methods and for proving many identities.

5.2 Derivative identities

Derivatives of Legendre polynomials can be expressed using related recurrence forms. One common identity relates \((1-x^2)P_n'(x)\) to a combination of \(P_{n-1}\) and \(P_n\), reflecting how differentiation interacts with the differential equation structure. Such formulas are frequently used in spherical-coordinate derivations and in evaluating surface flux terms.

5.3 Integral identities

Integrals involving \(P_n\) follow from orthogonality and recurrence. For example, certain antiderivatives can be written in terms of \(P_{n+1}\) and \(P_{n-1}\), and integrals of products can often be reduced to Kronecker-delta expressions using orthogonality. These identities help compute coefficients in expansions and evaluate moments.

5.4 Christoffel–Darboux-type relations (where applicable)

Legendre polynomials also satisfy summation formulas of the Christoffel–Darboux type, which express partial sums of products \(\sum_{k=0}^{n} P_k(x)P_k(y)\) in a closed form involving just \(P_n\) and \(P_{n+1}\). Such relations can accelerate computations and are useful in analyzing kernels associated with spectral expansions.

6 Spectral expansions and applications

6.1 Legendre series expansion of functions

Because \(\{P_n\}\) is an orthogonal basis on \([-1,1]\), a broad class of functions \(f(x)\) can be expanded as \[ f(x)\sim\sum_{n=0}^{\infty} a_n P_n(x), \] with coefficients determined by projection. In practice, truncating the series yields approximations that converge under typical regularity and integrability assumptions.

6.2 Coefficient formulas (projection onto polynomials)

Using orthogonality, the coefficients take the form \[ a_n=\frac{2n+1}{2}\int_{-1}^{1} f(x)\,P_n(x)\,dx. \] This formula is the Legendre analogue of Fourier coefficients and is widely used in numerical methods (spectral and pseudo-spectral schemes).

6.3 Convergence considerations

Convergence depends on the smoothness of \(f\). For sufficiently smooth functions, coefficients decay rapidly, and the truncated series approximates \(f\) well. If \(f\) has limited regularity or discontinuities, convergence may slow and oscillatory artifacts can occur near nonsmooth points, consistent with behavior seen in other orthogonal polynomial expansions.

6.4 Typical differential-equation separation contexts

Legendre’s equation emerges via separation of variables in spherical coordinates. In many settings, an angular dependence is governed by an eigenvalue equation that reduces to Legendre’s differential equation, with the eigenvalue matching \(\ell(\ell+1)\). The resulting polynomials (or their associated generalizations) provide the angular basis for solutions to PDEs like Laplace’s equation and related wave or heat problems under spherical symmetry.

7 Properties and symmetries

7.1 Parity: even/odd behavior

Legendre polynomials satisfy \[ P_n(-x)=(-1)^n P_n(x). \] Thus \(P_n\) is even when \(n\) is even and odd when \(n\) is odd. This parity is visible in the power-series solution and simplifies many integrals by symmetry.

7.2 Zeros and qualitative distribution

For \(n\ge 1\), \(P_n(x)\) has exactly \(n\) real zeros in the open interval \((-1,1)\). These roots interlace with those of neighboring degrees: between consecutive zeros of \(P_n\) there is at least one zero of \(P_{n+1}\) and vice versa. Such qualitative information is important in quadrature and stability considerations.

7.3 Asymptotic behavior (overview level)

For large degree \(n\), Legendre polynomials exhibit oscillatory behavior in \((-1,1)\) and different scaling near the endpoints. Qualitative asymptotics can be described using classical methods (e.g., stationary-phase-like arguments or known asymptotic expansions). These results help estimate truncation errors and guide numerical evaluation for high \(n\).

8 Computational aspects

8.1 Stable evaluation strategies

Direct evaluation using high-order derivatives is generally inefficient and numerically unstable. Instead, stable computation relies on recurrence relations and orthogonality-based methods. For large \(n\), careful handling is needed near \(x=\pm 1\) to avoid loss of significance.

8.2 Recurrence-based computation

The three-term recurrence allows sequential generation: start from \(P_0(x)\) and \(P_1(x)\), then compute \(P_{2}(x),P_{3}(x),\dots\). This approach is fast and often accurate when implemented with appropriate floating-point precautions. It also enables computing multiple degrees efficiently for spectral methods.

8.3 Numerical integration and orthogonality checks

When computing Legendre coefficients numerically, integrals on \([-1,1]\) are often evaluated using quadrature methods tailored to orthogonality, such as Gauss–Legendre quadrature. A practical verification is to check approximate orthogonality by integrating products \(P_m P_n\) and confirming that cross terms are close to zero within numerical tolerance.

9.1 Connection to spherical harmonics

Legendre’s equation forms the angular component underlying spherical harmonics. The full spherical-harmonic angular dependence involves both Legendre polynomials and associated functions, with eigenvalues corresponding to \(\ell(\ell+1)\). In many physical problems, regularity and boundary conditions lead directly to integer degree solutions that fit the spherical-harmonic framework.

9.2 Relation to associated Legendre equation (parameterized extension)

A common extension introduces an additional parameter (often denoted \(m\)), producing the associated Legendre equation. This generalization retains the same differential-operator structure but modifies the eigenvalue term and introduces additional singular behavior. The resulting associated functions arise in separations that include azimuthal dependence.

9.3 Comparison with Chebyshev and other classical orthogonal polynomials

Legendre polynomials are part of the broader family of classical orthogonal polynomials generated by Sturm–Liouville problems. Compared with Chebyshev polynomials, they use a different weight function and satisfy different recurrence relations, leading to distinct approximation properties. Nonetheless, they share many conceptual features with other orthogonal bases: orthogonality on an interval, recurrence-based computation, and applicability to spectral expansions of smooth functions.