Legendre Polynomial

DOWNLOAD Mathematica Notebook LegendreP

The Legendre polynomials, sometimes called Legendre functions of the first kind, Legendre coefficients, or zonal harmonics (Whittaker and Watson 1990, p. 302), are solutions to the Legendre differential equation. If l is an integer, they are polynomials. The Legendre polynomials P_n(x) are illustrated above for x in [-1,1] and n=1, 2, ..., 5. They are implemented in the Wolfram Language as LegendreP[n, x].

The associated Legendre polynomials P_l^m(x) and P_l^(-m) are solutions to the associated Legendre differential equation, where l is a positive integer and m=0, ..., l.

The Legendre polynomial P_n(z) can be defined by the contour integral

 P_n(z)=1/(2pii)∮(1-2tz+t^2)^(-1/2)t^(-n-1)dt,
(1)

where the contour encloses the origin and is traversed in a counterclockwise direction (Arfken 1985, p. 416).

The first few Legendre polynomials are

P_0(x)=1
(2)
P_1(x)=x
(3)
P_2(x)=1/2(3x^2-1)
(4)
P_3(x)=1/2(5x^3-3x)
(5)
P_4(x)=1/8(35x^4-30x^2+3)
(6)
P_5(x)=1/8(63x^5-70x^3+15x)
(7)
P_6(x)=1/(16)(231x^6-315x^4+105x^2-5).
(8)

When ordered from smallest to largest powers and with the denominators factored out, the triangle of nonzero coefficients is 1, 1, -1, 3, -3, 5, 3, -30, ... (OEIS A008316). The leading denominators are 1, 1, 2, 2, 8, 8, 16, 16, 128, 128, 256, 256, ... (OEIS A060818).

The first few powers in terms of Legendre polynomials are

x=P_1(x)
(9)
x^2=1/3[P_0(x)+2P_2(x)]
(10)
x^3=1/5[3P_1(x)+2P_3(x)]
(11)
x^4=1/(35)[7P_0(x)+20P_2(x)+8P_4(x)]
(12)
x^5=1/(63)[27P_1(x)+28P_3(x)+8P_5(x)]
(13)
x^6=1/(231)[33P_0(x)+110P_2(x)+72P_4(x)+16P_6(x)]
(14)

(OEIS A008317 and A001790). A closed form for these is given by

 x^n=sum_(l=n,n-2,...)((2l+1)n!)/(2^((n-l)/2)(1/2(n-l))!(l+n+1)!!)P_l(x)
(15)

(R. Schmied, pers. comm., Feb. 27, 2005). For Legendre polynomials and powers up to exponent 12, see Abramowitz and Stegun (1972, p. 798).

The Legendre polynomials can also be generated using Gram-Schmidt orthonormalization in the open interval (-1,1) with the weighting function 1.

P_0(x)=1
(16)
P_1(x)=[x-(int_(-1)^1xdx)/(int_(-1)^1dx)]·1
(17)
=x
(18)
P_2(x)=x[x-(int_(-1)^1x^3dx)/(int_(-1)^1x^2dx)]-[(int_(-1)^1x^2dx)/(int_(-1)^1dx)]·1
(19)
=x^2-1/3
(20)
P_3(x)=[x-(int_(-1)^1x(x^2-1/3)^2dx)/(int_(-1)^1(x^2-1/3)^2dx)](x^2-1/3)-[(int_(-1)^1(x^2-1/3)^2dx)/(int_(-1)^1x^2dx)]x
(21)
=x^3-3/5x.
(22)

Normalizing so that P_n(1)=1 gives the expected Legendre polynomials.

The "shifted" Legendre polynomials are a set of functions analogous to the Legendre polynomials, but defined on the interval (0, 1). They obey the orthogonality relationship

 int_0^1P^__m(x)P^__n(x)dx=1/(2n+1)delta_(mn).
(23)

The first few are

P^__0(x)=1
(24)
P^__1(x)=2x-1
(25)
P^__2(x)=6x^2-6x+1
(26)
P^__3(x)=20x^3-30x^2+12x-1.
(27)

The Legendre polynomials are orthogonal over (-1,1) with weighting function 1 and satisfy

 int_(-1)^1P_n(x)P_m(x)dx=2/(2n+1)delta_(mn),
(28)

where delta_(mn) is the Kronecker delta.

The Legendre polynomials are a special case of the Gegenbauer polynomials with alpha=1/2, a special case of the Jacobi polynomials P_n^((alpha,beta)) with alpha=beta=0, and can be written as a hypergeometric function using Murphy's formula

 P_n(x)=P_n^((0,0))(x)=_2F_1(-n,n+1;1;1/2(1-x))
(29)

(Bailey 1933; 1935, p. 101; Koekoek and Swarttouw 1998).

The Rodrigues representation provides the formula

 P_l(x)=1/(2^ll!)(d^l)/(dx^l)(x^2-1)^l,
(30)

which yields upon expansion

P_l(x)=1/(2^l)sum_(k=0)^(|_l/2_|)((-1)^k(2l-2k)!)/(k!(l-k)!(l-2k)!)x^(l-2k)
(31)
=1/(2^l)sum_(k=0)^(|_l/2_|)(-1)^k(l; k)(2l-2k; l)x^(l-2k)
(32)

where |_r_| is the floor function. Additional sum formulas include

P_l(x)=1/(2^l)sum_(k=0)^(l)(l; k)^2(x-1)^(l-k)(x+1)^k
(33)
=sum_(k=0)^(l)(l; k)(-l-1; k)((1-x)/2)^k
(34)

(Koepf 1998, p. 1). In terms of hypergeometric functions, these can be written

P_n(x)=((x-1)/2)^n_2F_1(-n,-n;1;(x+1)/(x-1))
(35)
P_n(x)=(2n; n)(x^n)/(2^n)_2F_1(-n/2,(1-n)/2;1/2-n;x^(-2))
(36)
P_n(x)=_2F_1(-n,n+1;1;(1-x)/2)
(37)

(Koepf 1998, p. 3).

A generating function for P_n(x) is given by

 g(t,x)=(1-2xt+t^2)^(-1/2)=sum_(n=0)^inftyP_n(x)t^n.
(38)

Take partialg/partialt,

 -1/2(1-2xt+t^2)^(-3/2)(-2x+2t)=sum_(n=0)^inftynP_n(x)t^(n-1).
(39)

Multiply (39) by 2t,

 -t(1-2xt+t^2)^(-3/2)(-2x+2t)=sum_(n=0)^infty2nP_n(x)t^n
(40)

and add (38) and (40),

 (1-2xt+t^2)^(-3/2)[(2xt-2t^2)+(1-2xt+t^2)]=sum_(n=0)^infty(2n+1)P_n(x)t^n
(41)

This expansion is useful in some physical problems, including expanding the Heyney-Greenstein phase function and computing the charge distribution on a sphere. Another generating function is given by

 sum_(n=0)^infty(P_n(x))/(n!)z^n=e^(xz)J_0(zsqrt(1-x^2)),
(42)

where J_0(x) is a zeroth order Bessel function of the first kind (Koepf 1998, p. 2).

The Legendre polynomials satisfy the recurrence relation

 (l+1)P_(l+1)(x)-(2l+1)xP_l(x)+lP_(l-1)(x)=0
(43)

(Koepf 1998, p. 2). In addition,

 (1-x^2)P_n^'(x)=-nxP_n(x)+nP_(n-1)(x)=(n+1)xP_n(x)-(n+1)P_(n+1)(x)
(44)

(correcting Hildebrand 1956, p. 324).

A complex generating function is

 P_l(x)=1/(2pii)int(1-2zx+z^2)^(-1/2)z^(-l-1)dz,
(45)

and the Schläfli integral is

 P_l(x)=((-1)^l)/(2^l)1/(2pii)int((1-z^2)^l)/((z-x)^(l+1))dz.
(46)

Integrals over the interval [x,1] include the general formula

 int_x^1P_m(x)dx=((1-x^2))/(m(m+1))(dP_m(x))/(dx)
(47)

for m!=0 (Byerly 1959, p. 172), from which the special case

int_0^1P_m(x)dx=(P_(m-1)(0)-P_(m+1)(0))/(2m+1)
(48)
={1 m=0; 0 m even !=0; (-1)^((m-1)/2)(m!!)/(m(m+1)(m-1)!!) m odd
(49)

follows (OEIS A002596 and A046161; Byerly 1959, p. 172). For the integral over a product of Legendre functions,

 int_x^1P_m(x)P_n(x)dx=((1-x^2)[P_n(x)P_m^'(x)-P_m(x)P_n^'(x)])/(m(m+1)-n(n+1))
(50)

for m!=n (Byerly 1959, p. 172), which gives the special case

 int_0^1P_m(x)P_n(x)dx={1/(2n+1)   m=n; 0   m!=n, m,n both even or odd; f_(m,n)   m even, n odd; f_(n,m)   m odd, n even
(51)

where

 f_(m,n)=((-1)^((m+n+1)/2)m!n!)/(2^(m+n-1)(m-n)(m+n+1)[(1/2m)!]^2{[1/2(n-1)]!}^2)
(52)

(OEIS A078297 and A078298; Byerly 1959, p. 172). The latter is a special case of

 int_0^1P_mu(x)P_nu(x)dx=(Asin(1/2pinu)cos(1/2pimu)-A^(-1)sin(1/2pimu)cos(1/2pinu))/(1/2pi(nu-mu)(mu+nu+1)),
(53)

where

 A=(Gamma(1/2(mu+1))Gamma(1+1/2nu))/(Gamma(1/2(nu+1))Gamma(1+1/2mu))
(54)

and Gamma(z) is a gamma function (Gradshteyn and Ryzhik 2000, p. 762, eqn. 7.113.1)

Integrals over [-1,1] with weighting functions x and x^2 are given by

int_(-1)^1xP_L(x)P_N(x)dx={(2(L+1))/((2L+1)(2L+3)) for N=L+1; (2L)/((2L-1)(2L+1)) for N=L-1
(55)
int_(-1)^1x^2P_L(x)P_N(x)dx={(2(L+1)(L+2))/((2L+1)(2L+3)(2L+5)) for N=L+2; (2(2L^2+2L-1))/((2L-1)(2L+1)(2L+3)) for N=L; (2L(L-1))/((2L-3)(2L-1)(2L+1)) for N=L-2
(56)

(Arfken 1985, p. 700).

The Laplace transform is given by

 L[P_n(t)](s)={1/2sqrt(pi)[sqrt(2/s)I_(-n-1/2)(s)-1/2s_1F_2(1;2+1/2n,1/2(3-n);1/4s^2)]   for n even; 1/2sqrt(pi)[sqrt(2/s)I_(-n-1/2)(s)+_1F_2(1;1/2(3+n),1-1/2n;1/4s^2)]   for n odd,
(57)

where I_n(s) is a modified Bessel function of the first kind.

A sum identity is given by

 1-[P_n(x)]^2=sum_(nu=1)^n(1-x^2)/(1-x_nu^2)[(P_n(x))/(P_n^'(x_nu)(x-x_nu))]^2,
(58)

where x_nu is the nuth root of P_n(x) (Szegö 1975, p. 348). A similar identity is

 sum_(nu=1)^n(1-x_nu^2)/((n+1)^2[P_(n+1)(x_nu)]^2)=1,
(59)

which is responsible for the fact that the sum of weights in Legendre-Gauss quadrature is always equal to 2.

Wolfram Web Resources

Mathematica »

The #1 tool for creating Demonstrations and anything technical.

Wolfram|Alpha »

Explore anything with the first computational knowledge engine.

Wolfram Demonstrations Project »

Explore thousands of free applications across science, mathematics, engineering, technology, business, art, finance, social sciences, and more.

Computerbasedmath.org »

Join the initiative for modernizing math education.

Online Integral Calculator »

Solve integrals with Wolfram|Alpha.

Step-by-step Solutions »

Walk through homework problems step-by-step from beginning to end. Hints help you try the next step on your own.

Wolfram Problem Generator »

Unlimited random practice problems and answers with built-in Step-by-step solutions. Practice online or make a printable study sheet.

Wolfram Education Portal »

Collection of teaching and learning tools built by Wolfram education experts: dynamic textbook, lesson plans, widgets, interactive Demonstrations, and more.

Wolfram Language »

Knowledge-based programming for everyone.