1

5.The Gamma Function (Factorial Function)

5.1 Definition, Simple Properties
At least three different, convenient definitions of the gamma function are
in common use
* Infinite limit.
The first definition, named after Euler, is

(5.1)


This definition is useful in developing the infinite-product form
of (z). Here z may be either real or complex. Replacing z with z+1, we have

(5.2)

, 3 , 2 , 1 , 0 ,
) ( ) 2 )( 1 (
3 2 1
lim ) ( =
+ + +

I

z n
n z z z z
n
z
z
n
). (
) ( ) 2 )( 1 (
3 2 1
1
lim
) 1 ( ) 3 )( 2 )( 1 (
3 2 1
lim ) 1 (
1
z z
n
n z z z z
n
n z
nz
n
n z z z z
n
z
z
n
z
n
I =
+ + +

+ +
=
+ + + + +

+ I

+

Appears in normalization of Coulomb wave functions and the computation of probabilities in statistical mechanics
2
1
) 1 ( 3 2 1
3 2 1
lim ) 1 ( =
+

= I

n
n n
n
n

)!. 1 ( ) 1 ( 3 2 1 ) (
, 2 ) 2 ( 2 ) 3 (
, 1 ) 2 (
= = I
= I = I
= I
n n n

* Definite integral (Euler)

This is the basic functional relation for G (z) . Also from the definition
(5.3)
(5.4)
A second definition, also frequently called Eulers form, is

, ) (
0
1
}

I dt t e z
z t

. 0 ) ( > 9 z
(5.5)
3
}

= I
0
1 2
, 2 ) (
2
dt t e z
z t
| | , ) ln( ) (
1
0
1
1
}

= I dt z
z
t
, 0 ) ( > 9 z
The restriction on z is necessary to avoid divergence of the integral. When G (z)
does appear in physical problems, it is often in this form or some variation such as
. 0 ) ( > 9 z
When z=1/2, Eq.(5.6) is just Gauss error function, and we have the interesting result
t = I ) (
2
1
(5.8)
(5.6)
(5.7)
To show the equivalence of these two definitions, consider the function of two
variables
}

=
n
z n
n
t
dt t n z F
0
1
, ) 1 ( ) , (
0 ) ( > 9 z
(5.9)
with n is a positive integer. Since
, ) 1 ( lim
t n
n
t
n
e

=
). ( ) , ( ) , ( lim
0
1
z dt t e z F n z F
z t
n
I = =
}

0 ) ( > 9 z
(5.11)
4
. ) 1 ( ) , (
1
0
1
}

= du u u n n z F
z n z
. ) 1 ( ) 1 (
) , (
1
0
1
1
0
}

+ = du u u
z
n
z
u
u
n
n z F
z n
z
n
z
by Eq.(5.)
Returning to F(z, n) , we evaluate it in successive integration by parts. For
convenience let u= t/n . Then
Integrating by parts, we obtain
(5.12)
(5.13)
Repeating this, we finally get
.
) ( ) 2 )( 1 (
3 2 1
) 1 ( ) 1 (
1 ) 1 (
) , (
1
0
1
z
n z z
n
n z z z z
n
du u
n z z z
n n
n n z F
+ + +

=
+ +

=
}
+

(5.14)
5
), ( ) , ( ) , ( lim z z F n z F
n
I =

, ) 1 (
) (
1
1
[

=

+
I
n
n z z
e
n
z
ze
z
This is identical with the expression on the right side of Eq.(5.1). Hence
(5.15)
by Eq.(5.1), completing the proof
The third definition is
(5.16)
where is the Euler-Mascheroni constant,
577216 . 0 ln
1
3
1
2
1
1 lim =
|
.
|

\
|
+ + + + =

n
n
n
(5.17)
* I nfinite product
This form can be derived from the original definition by writing it a
6
. 1
1
lim
) ( ) 1 (
3 2 1
lim ) (
1
1
z
n
m
n
z
n
n
m
z
z
n
n z z z
n
z

=


[
|
.
|

\
|
+ =
+ +

= I

,
) ln ( z n z
e n

=
, ) 1 ( lim
) (
1
1
) ln (
[
=

+ =
I
n
m
z n
n
m
z
e z
z
Inverting and using
we obtain
Multiplying and dividing by
, )
1
3
1
2
1
1 ( exp
1
[
=
=
(

+ + + +
n
m
m z
e z
n

we get
(5.18)
(5.19)
(5.20)
(5.21)
7
. ) 1 ( lim
ln
1
3
1
2
1
1 exp lim
) (
1
1
(

+
)
`

|
.
|

\
|
+ + + + =
I
[
=

n
m
m z
n
n
e
m
z
z n
n
z
z

[

+
I

1
) 1 (
) (
1
n
n z z
e
n
z
ze
z

, !
0
z dt t e
z t

* Factorial notation
Eq.(5.5) can be rewritten as
, 1 ) ( > 9 z
to define a factorial function z! The factorial function of Eq.(5.5) is, of course,
related to the gamma function by
(5.22)
(5.25)
8
!. ) 1 (
)! 1 ( ) (
z z
z z
= + I
= I
, 3 12 ! ! n n z = =
If z =n, a positive integer, we have
However, it should be noted carefully
that z! is now defined by (5.25) the
factorial function is no longer limited to
positive integer values of the argument
(Fig.5.1). The difference relation
(Eq.(5.2)) becomes
.
!
)! 1 (
z
z
z =
(5.27)
(5.28)
(5.29)
9
This shows immediately that
0!=1 (5.30)

for n, a negative integer.
* Double factorial notation
In many problems, we encounter products of the odd (or even) positive integers. For
convenience these are given special labels as double factorial:
! )! 2 ( ) 2 ( 6 4 2
! )! 1 2 ( ) 1 2 ( 5 3 1
n n
n n
=
+ = +

= ! n
(5.33b)
Clearly, these are related to the regular factorial by
! 2
)! 1 2 (
! )! 1 2 (
n
n
n
n
+
= +
(5.33c)
10
5.2 Digamma Functions
z
n
n
n z z z
n
z z z
) ( ) 2 )( 1 (
!
lim ) ( !
+ + +
= I =


| |, ) ln( ) 2 ln( ) 1 ln( ln ) ! ln( lim ) ! ln( n z z z n z n z
n
+ + + + =


As may be noted from the three definitions in the previous section, it is
inconvenient to deal with the derivatives of the gamma or factorial function directly.
Instead, it is customary to take the natural logarithm of the factorial
function(Eq.(5.1)), convert the product to a sum, and then differentiate, that is
Differentiating with respect to z, we obtain
),
1
2
1
1
1
(ln lim ) ( ) ! ln(
n z z z
n z F z
dz
d
n
+

+

+
=


(5.36)
(5.37)
(5.38)
11
which defines F(z), the digamma function. From the definition of the Euler constant,
the above equation may be rewritten as
.
) (
)
1 1
( ) (
1 1


=

=
+
+ =
+
=
n n
z n n
z
n n z
z F
5.3 The Beta Function
Using the integral definition(Eq.(5.25)), we write the product of two factorials as
the product of two integrals. To facilitate a change in variables, we take the integrals
over a finite range.
, lim ! !
2 2
2
0 0
} }


=
a a
n v m u
a
dv v e du u e n m
. 1 ) (
, 1 ) (
> 9
> 9
n
m
(5.39)
(5.57a)
Replacing u with x
2
and v with y
2
, we obtain
. 4 lim ! !
0 0
1 2 1 2
2 2
} }
+ +

=
a a
n y m x
a
dy y e dx x e n m
(5.57b)
12
Transforming to polar coordinates gives us
. sin cos 2 )! 1 (
sin cos 4 lim ! !
2
0
1 2 1 2
0
2
0
1 2 1 2 3 2 2
2
}
} }
+ +
+ + + +

+ + =
=
t
t
u u u
u u u
d n m
d dr r e n m
n m
a
n m n m r
a
The definite integral, together with the factor 2, has been named the beta function
). 1 , 1 (
)! 1 (
! !
sin cos 2 ) 1 , 1 (
2
0
1 2 1 2
+ + =
+ +
=
+ +
}
+ +
m n B
n m
n m
d n m B
n m
t
u u u
Equivalently, in terms of the gamma function
.
) (
) ( ) (
) , (
q p
q p
q p B
+ I
I I
=
(5.58)
(5.59a)
(5.59b)
13
* Definite integrals, alternative forms
The beta function is useful in the evaluation of a wide variety of definite
integrals. The substitution t=(cos)
2
converts Eq.(5.59) to
. ) 1 (
)! 1 (
! !
) 1 , 1 (
1
0
}
=
+ +
= + + dt t t
n m
n m
n m B
n m
Replacing t by x
2
, we obtain
. ) 1 (
)! 1 ( 2
! !
1
0
2 1 2
}
=
+ +
+
dx x x
n m
n m
n m
The substitution
) 1 ( u u t + = in Eq.(5.60a) yields still another useful form
.
) 1 ( )! 1 (
! !
0
2 }

+ +
+
=
+ +
du
u
u
n m
n m
n m
m
(5.60a)
(5.60b)
(5.61)
14
* Verification of a / sin a = a ! (- a )! relation
If we take m= a , n= -a , 0< a < 1, then
)!. ( !
) 1 (
0
2
a a du
u
u
a
=
+
}

On the other hand,

.
sin
1
) 1 ( 1 ) 1 (
0
1
0
0
2
a
a
e
dx e
a
u
du u
a
u
u
u
du u
x
ax
a a a
t
t
=
+
=
+
+
+

=
+
}
} }

) (
x
e u =
the last equality is obtained by using the previous result for the contour integral in
Chapter 2).
Therefore, we have proven the relation.
(5.62)
15
5.4 Incomplete Gamma functions
Generalizing the integral definition of the gamma function, we define
the incomplete gamma functions by the variable limit integrals
}

=
x
a t
dt t e x a
0
1
, ) , (
0 ) ( > 9 a
and
. ) . (
1
}

= I
x
a t
dt t e x a
Clearly, two functions are related, for
). ( ) , ( ) , ( a x a x a I = I +
) , ( x a
) , ( x a I The choice of employing or is purely a matter of
convenience. If the parameter a is a positive integer, Eq. (5.68) may be
integrated completely to yield
(5.68)
(5.69)
16
2 , 1 ,
!
)! 1 ( ) , (
!
1 )! 1 ( ) , (
1
0
1
0
= = I
|
|
.
|

\
|
=

n
s
x
e n x n
s
x
e n x n
n
s
s
x
n
s
s
x

* Error integrals
}
}

= =
=
z
t
z
t
dt e z erf z erfc
dt e z erf
2
2
2
) ( 1 ) (
2
) (
0
t
t
They can be written as incomplete gamma functions a =1/2 . The relations are
). , ( ) (
) , ( ) (
2
2
1
2 1
2
2
1
2 1
z z erf c
z z erf
I =
=

t
t

(5.70)
(5.8a)
(5.8b)