Добавил:
Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Скачиваний:
21
Добавлен:
08.01.2014
Размер:
988.22 Кб
Скачать

618 C Mathematical Formulas

ζ(3/2) = 2.612 ,

ζ(5/2)

= 1.341 .

(C.3.13)

When the sum runs over odd numbers only,

 

 

1

 

 

 

 

 

= (1 2−n)ζ(n) .

(C.3.14)

 

(2k + 1)n

k=0

On the other hand, when the terms are added with an alternating sign, the

Dirichlet eta function or alternating zeta function is obtained:

 

 

 

 

 

 

 

 

( 1)k−1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

η(n) =

 

 

 

 

= (1 21−n)ζ(n) .

(C.3.15)

 

 

 

 

kn

 

 

 

 

 

 

 

 

 

k=1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Its integral representation can be analytically continued as

 

1

 

z

1

 

 

 

 

 

 

 

1

 

 

 

 

 

 

 

 

 

 

 

 

x

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(1)k−1e−kxxz−1 dx .

 

η(z) =

 

 

 

dx =

 

 

 

k=1

(C.3.16)

Γ (z)

ex + 1

Γ (z)

0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0

 

 

 

 

 

 

 

 

 

 

 

For particular values of the argument

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

η(1) = ln 2 ,

 

 

η(2) =

π2

 

 

η(4) =

 

7π4

(C.3.17)

 

 

 

 

 

 

,

 

 

 

 

.

 

 

 

 

12

720

The Bernoulli Numbers and their Relationship with the Zeta

Function

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The Bernoulli numbers Bn are defined via the series expansion

 

 

 

 

 

 

 

 

 

 

 

 

 

z

 

 

=

 

 

 

 

zn

 

 

 

 

 

(C.3.18)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Bn

 

.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ez

1

 

 

 

 

 

n!

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n=0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The first few Bernoulli numbers are

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1

 

 

 

 

 

 

 

 

 

1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

B0 = 1 , B1 =

 

 

, B2 =

 

 

, B2n+1 = 0 ,

 

2

6

(C.3.19)

1

 

 

 

 

 

 

1

 

 

 

 

 

 

 

 

 

 

 

 

 

1

 

 

 

5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

B4 =

 

 

, B6 =

 

 

, B8 =

 

, B10

=

 

.

 

30

42

30

66

 

For positive even (and negative odd) numbers the ζ function can also be

expressed in terms of the Bernoulli numbers:

 

ζ(2n) = (1)n−1

(2π)2n

(C.3.20)

 

 

 

 

B2n

2(2n)!

and

 

 

B2n

 

ζ(1 2n) =

(C.3.21)

 

 

,

2n

or conversely

(1)n−12(2n)!

 

 

 

 

B2n =

ζ(2n) .

(C.3.22)

 

(2π)2n

 

C.3 Special Functions

619

The Gamma Function

The gamma function (also called Euler’s gamma function or the factorial function) can be defined via the recursion relation

Γ (z + 1) = (z) ,

(C.3.23)

with the additional condition Γ (1) = 1. In the Re z > 0 region, when the condition Re k > 0 is met, it can be represented by the integrals

Γ (z) = 0

tz−1e−t dt = kz 0

tz−1e−kt dt .

(C.3.24)

It can be easily shown by integration by parts that these expressions satisfy the above recursion formula. Via analytical continuation they can be extended to the whole complex plane, where it has poles at nonpositive integers.

By iterating the recursion formula for integer values of z,

Γ (n) = (n − 1)! .

(C.3.25)

Therefore the factorial can be defined for noninteger values of z via the gamma function. For z = 12

Γ ( 21 ) = 0

t1/2e−t dt = 2

0

e−t2

dt =

 

,

(C.3.26)

π

thus for half-integer values of z

 

 

 

Γ (n + 1 ) =

(2n − 1)!!

 

.

 

π

 

2

 

 

 

2n

 

 

 

 

 

 

 

Consequently 21 ! = Γ ( 23 ) = 21

 

= 0.886.

 

 

 

 

 

 

 

 

π

 

 

 

 

 

 

 

For large values of the argument,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Γ (x) = /

 

x

xx e−x 1 +

 

12x

+ 288x2 + . . . .

 

 

2π

1

1

 

 

 

 

 

This leads to the Stirling formula:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n

 

 

n

 

 

 

 

n! = (n)

 

 

2πn n

e

.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The digamma function (or psi function) is defined as

ψ(z) = ddz ln Γ (z) .

(C.3.27)

(C.3.28)

(C.3.29)

(C.3.30)

Apart from negative integers, the function can be given in the form of an asymptotic series,

620 C Mathematical Formulas

 

 

 

1

1

 

(C.3.31)

ψ(z + 1) = −γ − n=1 z + n

n ,

 

 

 

 

 

 

 

 

where γ is the Euler–Mascheroni constant (or Euler’s constant):

 

 

n

1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

γ = −ψ(1) = n→∞ #

k ln n$

0.577 215 . . . .

(C.3.32)

lim

k=1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

For z = 1

 

 

 

 

 

 

 

 

2

 

 

 

 

 

 

 

 

ψ( 21 ) = −γ − 2 ln 2 .

 

 

 

(C.3.33)

C.3.3 Bessel Functions

The Bessel functions (also called cylinder functions) of order ν are defined as solutions to the Bessel di erential equation

z2

d2w(z)

+ z

dw(z)

+ (z2 − ν2)w(z) = 0 ,

(C.3.34)

dz2

 

dz

where ν can be any real or complex number. One particular solution can be written as a power series of z,

( 1)k

 

z

 

2k+ν

(C.3.35)

Jν (z) = k=0 k!Γ (ν+ k + 1)

2

.

 

 

 

 

 

 

 

 

Except for integer orders ν = n, when

 

 

 

 

 

J−n(z) = (1)nJn(z) ,

 

 

(C.3.36)

Jν (z) and J−ν (z) are linearly independent, and the general solution of the Bessel di erential equation is

Zν (z) = c1Jν (z) + c2J−ν (z) .

(C.3.37)

There are two classes of solutions to the Bessel equation for any ν with di erent analytic properties. The functions Jν (z) are called Bessel functions of the first kind (or simply Bessel functions), while the Bessel functions of the second kind (more commonly called Neumann functions or Weber functions) can be written as

Nν (z) =

Jν (z) cos(νπ) − J−ν (z)

.

(C.3.38)

 

sin(νπ)

 

For integer orders n the Neumann function Nn(z) is defined as the ν → n limit of Nν (z). The Bessel functions of the third kind, also known as Hankel functions are special combinations of the Bessel functions of the first and second kinds:

 

 

 

 

 

 

C.3 Special Functions 621

Hν(1)(z) = Jν (z) + iNν (z) ,

 

 

Hν(2)(z) = Jν (z) iNν (z) .

(C.3.39)

For large values of |z|

 

 

 

 

 

 

 

 

 

 

Jν (z) /

πz

cos

z − 21 πν − 41 π ,

(C.3.40-a)

 

 

2

 

 

 

 

 

 

 

 

 

 

 

 

Nν (z) /

2

 

 

21 πν − 41 π

(C.3.40-b)

πz sin

z −

 

 

 

 

 

 

 

 

asymptotically.

The Bessel functions of integer order have a simple integral representation:

 

i−n

π

 

i−n

2π

 

 

Jn(z) =

0

eiz cos θ cos() dθ =

0

eiz cos θ eidθ .

(C.3.41)

π

2π

Among the Bessel functions of fractional order particularly important are the Bessel functions of half-integer order. Instead of Jn+1/2(z), it is customary

to use the spherical Bessel function of the first kind

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

jn(z) = /

 

π

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Jn+1/2(z) ,

(C.3.42)

 

 

 

 

2z

which satisfies the equation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

d2jn(z)

 

djn(z)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

z2

 

 

+ 2z

 

 

 

+ [z2 − n(n + 1)]jn(z) = 0 ,

(C.3.43)

dz2

dz

 

 

where n = 0, ±1, ±2, . . . . It can be shown that

 

 

 

 

jn(z) = zn

1 d

n sin z

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.

(C.3.44)

 

 

 

z

dz

 

z

The explicit expressions for the first few spherical Bessel functions are

 

 

 

j0(z) =

sin z

,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

z

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

j1(z) =

sin z

cos z

 

 

 

 

 

 

(C.3.45)

 

 

 

 

 

 

 

 

 

,

 

 

 

 

 

 

 

 

 

 

z2

 

z

 

 

 

 

 

 

 

 

 

 

 

 

 

3

 

1

 

 

 

 

 

 

 

 

3

 

 

 

 

 

 

j2(z) =

 

 

 

 

sin z −

 

cos z .

 

 

 

 

z3

z

z2

 

For small values of z the leading-order contribution is

 

 

 

 

 

jn(z)

 

 

 

 

 

 

zn

 

,

(C.3.46)

 

 

 

 

(2n + 1)!!

while for large values of z

622 C Mathematical Formulas

1

cos

 

i(n+1)

eiz + (1)n+1eiz

 

jn(z)

 

z − 21 (n + 1)π

 

(C.3.47)

z

2z

asymptotically.

Just like the Bessel di erential equation, (C.3.43) also has other sets of solutions, the spherical Neumann functions or spherical Bessel functions of the second kind :

nn(z) =

/

 

 

 

/

 

 

 

(C.3.48)

 

2z Nn+1/2(z) = (1)n+1

 

2z J−n−1/2(z) ,

 

 

 

π

 

 

π

 

and the spherical Hankel functions or spherical Bessel functions of the third kind :

h(1)n (z) = jn(z) + inn(z) , h(2)n (z) = jn(z) inn(z) . (C.3.49)

In analogy with (C.3.44), the spherical Bessel functions of the second kind may be written as

nn(z) = −zn

1 d

 

n cos z

(C.3.50)

 

 

 

 

 

.

z

dz

z

The first few spherical Bessel functions of the second and third kinds are

n0(z) =

cos z

,

 

 

 

z

 

 

 

n1(z) =

cos z

sin z

,

(C.3.51-a)

z2

z

n2(z) =

3

1

3

sin z ,

 

 

cos z −

 

z3

z

z2

h(1)0 (z) = zi eiz ,

 

 

 

 

 

 

i

 

 

(1)

(z) = eiz

1

 

 

(C.3.51-b)

h1

z

z2

,

h2

(z) = eiz

z z2 z3 .

 

(1)

 

 

i

3

 

 

3i

 

For small values of z the leading-order singularity of nn(z) is of the type

 

n

n

(z)

 

 

(2n − 1)!!

,

 

(C.3.52)

 

 

 

 

 

 

 

 

zn+1

 

 

while for large values of z

 

 

 

 

 

 

 

 

 

 

 

 

 

 

nn(z)

 

 

 

1

sin

 

z

 

1 (n + 1)π

,

(C.3.53-a)

 

 

 

 

 

 

 

 

 

z

 

 

2

 

 

 

h(1)

(z)

1

 

 

 

 

 

 

 

ei[z−(n+1)π/2]

 

 

(C.3.53-b)

n

 

z

 

 

 

 

 

 

 

 

asymptotically.

C.4 Orthogonal Polynomials

623

The modified Bessel functions of order ν satisfy the modified Bessel differential equation

z2 d2w(z) dz2

+ z

dw(z)

(z2 + ν2)w(z) = 0 .

(C.3.54)

dz

Comparison with the Bessel di erential equation immediately shows that the series expansion in powers of z,

1

 

 

z

 

ν+2k

Iν (z) = k=0 k!Γ (ν + k + 1)

2

(C.3.55)

 

 

 

 

 

 

 

is a solution. This is called modified Bessel function of the first kind. For integer orders ν = n

In(z) = i−nJn(iz) ,

(C.3.56)

while for general ν the solutions Iν (z) and I−ν (z) are linearly independent. A particular linear combination of these functions,

Kν (z) =

π

 

I−ν (z) − Iν (z)

(C.3.57)

 

sin(νz)

2

 

 

is the modified Bessel function of the second kind or Macdonald function. It can also be obtained from the analytical continuation of the Hankel function:

Kν (z) = 1 iπeiπν/2H(1)

(iz) .

(C.3.58)

2

ν

 

 

Of special importance is the function large values of its argument are

ln(

K0(z) = / π e

2z

K0; its asymptotic form for small and

12 z) + γ

z

1

,

−z

z

 

1

(C.3.59)

 

,

 

 

 

 

where γ = 0.57721 is the Euler–Mascheroni constant given in (C.3.32).

C.4 Orthogonal Polynomials

C.4.1 Hermite Polynomials

The Hermite polynomials are the solutions of the di erential equation

d2Hn(x)

2x

dHn(x)

+ 2nHn(x) = 0 , n = 0, 1, 2, . . . .

(C.4.1)

dx2

dx

A compact representation is given by Rodrigues’ formula:

624 C Mathematical Formulas

Hn(x) = (1)n ex2 dn e−x2 dxn

They are orthogonal with respect to the weight function w(x) = e−x

e−x2 Hn(x)Hm(x) dx = 2nn!πδnm .

−∞

(C.4.2)

2:

(C.4.3)

The explicit expressions for the first few Hermite polynomials are

H0(x) = 1 , H1(x) = 2x ,

H2(x) = 4x2 1 , (C.4.4)

H3(x) = 8x3 12x ,

H4(x) = 16x4 48x2 + 12 , while their general formula is

[n/2]

 

n!(2x)n−2k

 

 

 

 

 

(1)k k!(n

2k)! .

(C.4.5)

Hn(x) =

 

k=0

C.4.2 Laguerre Polynomials

The Laguerre polynomials are the solutions of the di erential equation

x

d2Ln(x)

+ (1 − x)

dLn(x)

+ nLn(x) = 0

 

n = 0, 1, 2, . . . .

(C.4.6)

dx2

 

dx

 

Alternatively, they can be given by Rodrigues’ formula:

 

 

 

 

 

 

1

ex

dn

 

e−xxn .

 

 

 

 

Ln(x) =

 

 

(C.4.7)

 

 

 

n!

dxn

They are orthogonal with respect to the weight function w(x) = e−x:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0

e−xLm(x)Ln(x) dx = δm,n .

(C.4.8)

The explicit expressions for the first few Laguerre polynomials are

 

 

 

L0(x) = 1 ,

 

 

 

 

 

 

 

 

 

 

L1(x) = 1 − x ,

1

 

 

 

 

 

 

L2(x) = 1 2x +

x2 ,

 

 

(C.4.9)

 

 

 

 

 

 

 

2

 

 

 

 

L3(x) = 1 3x +

3

x2

1

x3 .

 

 

 

2

6

 

C.4 Orthogonal Polynomials

625

The generalized Laguerre polynomials are the solutions of the di erential equation

x

d2Ln(α)(x)

+ (α + 1 − x)

dLn(α)(x)

+ nLn(α)(x) = 0

(C.4.10)

dx2

dx

 

for nonnegative values of n, where α is an arbitrary complex number. Consequently, Rodrigues’ formula reads

 

1

 

dn

 

Ln(α)(x) =

 

exx−α

 

e−xxn+α .

(C.4.11)

n!

dxn

For α > −1 the generalized Laguerre polynomials satisfy the orthogonality relation with respect to the weight function w(x) = e−xxα

e−xxαLn(α)(x)Ln

(x) dx =

n!

δnn .

(C.4.12)

0

 

(α)

 

Γ (α + n + 1)

 

 

Of particular interest are the generalized Laguerre polynomials whose index α is a nonnegative integer. Their polynomial form is

n

n + m

 

xk

 

 

 

 

Lnm(x) = k=0(1)k n

 

k

 

 

,

(C.4.13)

 

 

 

 

 

 

 

while their connection with ordinary Laguerre polynomials is given by

dm

 

Lnm(x) = (1)m dxm Ln+m(x)

(C.4.14)

and L0n(x) = Ln(x).

C.4.3 Legendre Polynomials

The Legendre polynomials are defined on the interval |x| ≤ 1, and satisfy the di erential equation

(1 − x2)

d2Pl(x)

2x

dPl(x)

+ l(l + 1)Pl(x) = 0 ,

(C.4.15)

 

dx2

 

dx

that is,

 

 

 

 

 

 

 

 

 

 

 

 

 

Pl(x) =

1 dl

 

1 [l/2]

l

2l

l

2k

(C.4.16)

2ll! dxl (x2 1)l =

2l

k=0(1)k k

 

xl−2k .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The Legendre polynomials are mutually orthogonal, however they are not normalized:

626 C Mathematical Formulas

+1

2l + 1 .

(C.4.17)

 

Pl(x)Pl (x) dx = δl,l

 

 

2

 

 

1

 

 

 

 

The explicit expressions of the first few polynomials are

P0(x) = 1 ,

 

 

 

 

 

 

 

 

P1

(x) = x = cos θ ,

 

 

 

 

 

 

 

 

1

(3x2

1

 

(3 cos 2θ + 1) ,

 

P2

(x) =

 

1) =

 

 

(C.4.18)

2

4

P3

(x) =

1

(5x3

3x) =

1

(5 cos 3θ + 3 cos θ) ,

 

2

 

8

 

 

 

1

 

 

 

1

 

 

P4

(x) =

 

(35x4 30x2 + 3) =

 

(35 cos 4θ + 20 cos 2θ + 9) .

 

8

64

 

The same equation, (C.4.15), is satisfied by the Legendre polynomials of the second kind,

Q0(x) =

Q1(x) =

Q2(x) =

1 ln

 

1 + x

 

 

,

 

 

 

 

1 − x

 

 

 

 

2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

x

 

1

+ x

1 ,

 

 

 

ln

 

 

 

(C.4.19)

2

1

− x

3x2

 

 

1

 

 

 

1 + x

 

3

 

 

 

ln

 

 

x .

 

4

 

1 − x

2

The associated Legendre polynomials are solutions of the equation

(1 − x )

d2P m(x)

 

dP m(x)

+

l(l + 1)

m2

Pl

(x) = 0 ,

(C.4.20)

dx2

2x

dx

1 − x2

2

l

 

 

l

 

 

 

 

m

 

that is

 

 

 

 

 

 

 

 

 

 

 

Plm(x) = (1 − x2)m/2

dmPl(x)

l, m = 0, 1, 2, . . .

m ≤ l ,

(C.4.21)

dxm

 

 

while the associated Legendre polynomials of the second kind are defined as

Qlm(x) = (1 − x2)m/2

dmQl(x)

l, m = 0, 1, 2, . . . m ≤ l . (C.4.22)

dxm

In several references an additional factor (1)m appears in the definition. The associated Legendre polynomials satisfy the orthogonality relation

+1

2l + 1 (l − m)! .

(C.4.23)

 

Plm(x)Plm(x) dx = δl,l

 

 

2

 

(l + m)!

 

1

 

 

 

 

 

 

The explicit expressions for the first few associated Legendre polynomials are

 

 

 

 

 

 

 

 

 

 

 

C.4

Orthogonal Polynomials 627

P11(x) = (1 − x2)1/2 = sin θ ,

 

 

 

P21

(x) = 3(1 − x2)1/2x =

3

sin 2θ ,

 

 

 

 

2

 

 

 

 

3

(1

cos 2θ) ,

 

P22

(x) = 3(1 − x2) =

 

(C.4.24)

2

 

3

(1 − x2)1/2(5x2 1) =

3

 

(sin θ + 5 sin 3θ) ,

P31

(x) =

 

 

 

2

8

 

 

15

(cos θ − cos 3θ) ,

P32

(x) = 15(1 − x2)x =

 

 

4

P33

(x) = 15(1 − x2)3/2 =

15

(3 sin θ − sin 3θ) .

 

4

C.4.4 Spherical Harmonics

In physics, instead of the associated Legendre polynomials, the functions

 

 

 

m+|m|

 

1

 

1/2

 

(2l + 1)(l

m )!

 

1/2

 

m

l

 

 

 

l

 

 

2π

 

2(l + |m|)!

 

 

Y m(θ, ϕ) = (

 

1) 2

 

 

 

 

 

− |

|

 

 

P | |(cos θ)ei

 

 

 

 

 

 

 

 

 

 

(C.4.25) are used, where −l ≤ m ≤ l. They are called spherical harmonics (or surface harmonics). Spherical harmonics satisfy the spherical harmonic di erential equation, which is given by the angular part of Laplace’s equation in spherical coordinates. It is immediately seen from the above form that

−m(θ, ϕ) = ( 1)mY m (θ, ϕ) .

(C.4.26)

Yl

l

The spherical harmonics are normalized in such a way that the orthogonality relation

2π π

Y

m

m

(θ, ϕ) sin θ dθ dϕ = δll δmm

(C.4.27)

l

(θ, ϕ)Y

l

 

 

 

 

0 0

and the completeness relation

+l

 

δ(θ − θ )δ(ϕ − ϕ )

(C.4.28)

Y m (θ, ϕ)Y m(θ , ϕ ) =

 

 

 

 

 

l

l

 

 

 

l=0 m=−l

 

sin θ

 

 

 

 

 

= δ(cos θ − cos θ )δ(ϕ − ϕ ) ≡ δ(Ω − Ω )

are both satisfied.

According to the addition theorem for spherical harmonics, the product of two spherical harmonics can be written as the linear combination of spherical harmonics:

Yl1

1 (θ, ϕ)Yl2

2 (θ, ϕ) =

7

 

4π(2l + 1)

(C.4.29)

 

 

l1+l2

l

 

(2l1 + 1)(2l2 + 1)

 

m

m

|

 

 

 

 

 

 

 

 

 

 

l= l1 −l2| m=−l

 

 

 

 

×(l1l2m1m2|lm)(l1l200|l0)Ylm(θ, ϕ) .