ProofComplete

Special Functions - Key Proof

We prove the orthogonality of Legendre polynomials using Sturm-Liouville theory, demonstrating the general method applicable to all classical special functions.

ProofOrthogonality of Legendre Polynomials

Theorem: The Legendre polynomials Pm(x)P_m(x) and Pn(x)P_n(x) satisfy:

11Pm(x)Pn(x)dx=22n+1δmn\int_{-1}^1 P_m(x)P_n(x)dx = \frac{2}{2n+1}\delta_{mn}

Proof:

Step 1: Legendre polynomials satisfy Legendre's equation:

ddx[(1x2)dPndx]+n(n+1)Pn=0\frac{d}{dx}\left[(1-x^2)\frac{dP_n}{dx}\right] + n(n+1)P_n = 0

This is a Sturm-Liouville problem with p(x)=1x2p(x) = 1-x^2, q(x)=0q(x) = 0, w(x)=1w(x) = 1, and eigenvalues λn=n(n+1)\lambda_n = n(n+1).

Step 2: Write the equation for PmP_m and PnP_n:

ddx[(1x2)dPmdx]+m(m+1)Pm=0\frac{d}{dx}\left[(1-x^2)\frac{dP_m}{dx}\right] + m(m+1)P_m = 0

ddx[(1x2)dPndx]+n(n+1)Pn=0\frac{d}{dx}\left[(1-x^2)\frac{dP_n}{dx}\right] + n(n+1)P_n = 0

Step 3: Multiply the first equation by PnP_n and the second by PmP_m:

Pnddx[(1x2)dPmdx]+m(m+1)PmPn=0P_n\frac{d}{dx}\left[(1-x^2)\frac{dP_m}{dx}\right] + m(m+1)P_m P_n = 0

Pmddx[(1x2)dPndx]+n(n+1)PmPn=0P_m\frac{d}{dx}\left[(1-x^2)\frac{dP_n}{dx}\right] + n(n+1)P_m P_n = 0

Step 4: Subtract the second from the first:

Pnddx[(1x2)dPmdx]Pmddx[(1x2)dPndx]+[m(m+1)n(n+1)]PmPn=0P_n\frac{d}{dx}\left[(1-x^2)\frac{dP_m}{dx}\right] - P_m\frac{d}{dx}\left[(1-x^2)\frac{dP_n}{dx}\right] + [m(m+1) - n(n+1)]P_m P_n = 0

Step 5: The first two terms combine as:

ddx[(1x2)(PndPmdxPmdPndx)]\frac{d}{dx}\left[(1-x^2)\left(P_n\frac{dP_m}{dx} - P_m\frac{dP_n}{dx}\right)\right]

Thus:

ddx[(1x2)(PndPmdxPmdPndx)]=[n(n+1)m(m+1)]PmPn\frac{d}{dx}\left[(1-x^2)\left(P_n\frac{dP_m}{dx} - P_m\frac{dP_n}{dx}\right)\right] = [n(n+1) - m(m+1)]P_m P_n

Step 6: Integrate from 1-1 to 11:

[(1x2)(PndPmdxPmdPndx)]11=[n(n+1)m(m+1)]11PmPndx\left[(1-x^2)\left(P_n\frac{dP_m}{dx} - P_m\frac{dP_n}{dx}\right)\right]_{-1}^1 = [n(n+1) - m(m+1)]\int_{-1}^1 P_m P_n dx

The left side vanishes because (1x2)=0(1-x^2) = 0 at both x=±1x = \pm 1.

Step 7: For mnm \neq n, we have n(n+1)m(m+1)0n(n+1) - m(m+1) \neq 0, so:

11Pm(x)Pn(x)dx=0(mn)\int_{-1}^1 P_m(x)P_n(x)dx = 0 \quad (m \neq n)

Step 8: For normalization (m=nm = n), use Rodrigues' formula:

Pn(x)=12nn!dndxn(x21)nP_n(x) = \frac{1}{2^n n!}\frac{d^n}{dx^n}(x^2-1)^n

Compute:

11Pn2(x)dx=1(2nn!)211dndxn(x21)ndndxn(x21)ndx\int_{-1}^1 P_n^2(x)dx = \frac{1}{(2^n n!)^2}\int_{-1}^1 \frac{d^n}{dx^n}(x^2-1)^n \cdot \frac{d^n}{dx^n}(x^2-1)^n dx

Integrate by parts nn times (boundary terms vanish):

=(1)n(2nn!)211(x21)nd2ndx2n(x21)ndx= \frac{(-1)^n}{(2^n n!)^2}\int_{-1}^1 (x^2-1)^n \cdot \frac{d^{2n}}{dx^{2n}}(x^2-1)^n dx

Since (x21)n(x^2-1)^n is a polynomial of degree 2n2n, its 2n2n-th derivative is (2n)!(2n)!:

=(2n)!(2nn!)211(x21)ndx= \frac{(2n)!}{(2^n n!)^2}\int_{-1}^1 (x^2-1)^n dx

Evaluating this integral (using the beta function):

11(1x2)ndx=201(1x2)ndx=2n+1(n!)2(2n+1)!\int_{-1}^1 (1-x^2)^n dx = 2\int_0^1 (1-x^2)^n dx = \frac{2^{n+1}(n!)^2}{(2n+1)!}

Therefore:

11Pn2(x)dx=(2n)!(2nn!)22n+1(n!)2(2n+1)!=22n+1\int_{-1}^1 P_n^2(x)dx = \frac{(2n)!}{(2^n n!)^2} \cdot \frac{2^{n+1}(n!)^2}{(2n+1)!} = \frac{2}{2n+1}

RemarkGeneral Sturm-Liouville Orthogonality

This proof method applies to any Sturm-Liouville eigenfunctions. The key steps are:

  1. Write the differential equation for two eigenfunctions
  2. Form the combination that produces a total derivative
  3. Integrate and use boundary conditions
  4. Eigenvalue difference ensures orthogonality

For Bessel functions, Hermite polynomials, etc., the same technique proves orthogonality with respect to appropriate weight functions.

ProofCompleteness Sketch

Theorem: Any function f(x)L2([1,1])f(x) \in L^2([-1,1]) can be expanded as:

f(x)=n=0cnPn(x),cn=2n+1211f(x)Pn(x)dxf(x) = \sum_{n=0}^{\infty}c_n P_n(x), \quad c_n = \frac{2n+1}{2}\int_{-1}^1 f(x)P_n(x)dx

Proof Sketch:

The Legendre polynomials form a complete orthogonal basis because:

  1. They are eigenfunctions of a self-adjoint operator
  2. The Weierstrass approximation theorem shows polynomials are dense in C([1,1])C([1,1])
  3. Gram-Schmidt applied to monomials {1,x,x2,}\{1, x, x^2, \ldots\} produces the Legendre polynomials
  4. Since finite linear combinations of PnP_n can approximate any continuous function, the PnP_n span the space

Completeness guarantees:

n=0cn222n+1=11f(x)2dx\sum_{n=0}^{\infty}|c_n|^2\frac{2}{2n+1} = \int_{-1}^1 |f(x)|^2dx

(Parseval's theorem for Legendre expansion).

ExampleExpansion of a Step Function

For f(x)={10<x<111<x<0f(x) = \begin{cases}1 & 0 < x < 1 \\ -1 & -1 < x < 0\end{cases}:

cn=2n+1211f(x)Pn(x)dx=2n+1201Pn(x)dx2n+1210Pn(x)dxc_n = \frac{2n+1}{2}\int_{-1}^1 f(x)P_n(x)dx = \frac{2n+1}{2}\int_0^1 P_n(x)dx - \frac{2n+1}{2}\int_{-1}^0 P_n(x)dx

For odd nn: By symmetry Pn(x)=Pn(x)P_n(-x) = -P_n(x), so cn0c_n \neq 0 only for odd nn.

For even nn: Pn(x)=Pn(x)P_n(-x) = P_n(x), so integrals cancel and cn=0c_n = 0.

This gives:

f(x)=32P1(x)+78P3(x)+1116P5(x)+f(x) = \frac{3}{2}P_1(x) + \frac{7}{8}P_3(x) + \frac{11}{16}P_5(x) + \cdots

These proofs establish the rigorous mathematical foundation for using special functions as complete orthogonal bases in solving physical problems.