Complex Analysis Four 2
Complex Analysis Four 2
Complex Analysis Four 2
Tibra Ali
August 14, 2014
Lecture 4
z1
z2
zK+1
z3
zk
z4 zk+2
zk+3
I k
f (z)dz = 2i Res of f (z) at z = zi
C i =1
1
2 Examples
2.1 Trigonometric Integrals
First we consider integrals of the form
Z 2
1
I= d, with a > b > 0 and ( a, b) R.
0 a + b cos
Since this is an integral on the circle we can choose our contour C to be the unit circle on the complex plane.
We set
z = ei
which implies
idz
= d.
z
However, we want as our integrand to be a meromorphic function whose residues are easy to compute. So
let us choose
z + 1/z
cos = .
2
Then our integral becomes
dz
I
I = 2i .
C 2za + b + bz2
So we have to find the zeros of the polynomial
2
2.2 Semi-circular Contours
Before we evaluate integrals that use semi-circular contours let us state a useful theorem:
M
| f (z)|
Rk
on a contour parametrized by
z = Rei
for either 0 or 2 and, k > 1 and M are real positive constants. Then
Z
f (z)dz M
(1)
R k 1
and this vanishes as R for k > 1.
2.3 An Application
Integrals of the form
Z
dx
0 ( a2 + x2 )(b2 + x2 )
can be evaluated by using semi-circular contours centred about the origin. Because the integrand is an
R R
1 1
2 ( a + z )(b2 + z2 )
2 2
3
has poles in the upper half-plane at
z = ia and ib.
In the limit R the contour integral reduces to real integral that we want to evaluate since the contribu-
tion from the semicircular part goes to zero very rapidly. Thus we have
1 dz
Z
I=
2 (z2 + a2 )(z2 + b2 )
1 1 1
= 2i lim + lim
2 zia ( z + ia )( z2 + b2 ) zib ( z2 + a2 )( z + ib )
1
= .
2ab ( a + b)
In the above calculation we assumed that a 6= b. In case a = b we have double poles at z = ia and z = ia.
In this case the integral can be evaluated using the formula for higher order pole that we saw in the last
lecture:
1 d 1
2i lim (z ia)2 = 3.
2 zia dz (z ia)2 (z + ia)2 4a
Note that, care should be taken in taking the limit R . In doing so we are evaluating the Cauchy
principle value of the integral I. If the two sides of the semicircle were taken to infinity in different ways
we would get different results.
Another thing to note is that in the examples above we closed the contour in the upper half-plane, but
we could have easily chosen the contour to close in the lower-half plane and we would have been able
to evaluate the integral. This is because in the Cauchy principle value limit both contours reduce to the
integral that we want to evaluate. In the next example we will see that we have to choose one particular
contour.
4
2.5 Mousehole Contours
Mousehole contours are useful for evaluating integrals where the integrand has a simple pole at the origin.
Before we consider a concrete example of an integral let us prove a simple and useful theorem:
\beta \alpha
\epsilon
Suppose f (z) has a simple pole at z = a with residue then consider the contour
e = a + eeit , t ,
then
Z
lim f (z)dz = ( )i.
e0 e
By continutiy the contribution from the g(z) term is zero and the we have the desired result.
2.6 Example
Consider the integral
Z Z
x sin x 1 x sin x
dx = dx.
0 x3 2 x3
Now consider the function
iz eiz
f (z) = (2)
z3
z2
1
= 3 iz 1 iz + (3)
z 2!
The imaginary part f (z) on the real line coincides with our integrand but f (z) has a triple pole at z = 0. So
let us consider instead
iz eiz 1
h(z) = 3
+ 3.
z z
In defineing h(z) we have kept the imaginary part of the function on the real line unchanged but we have
subtracted the triple pole at z = 0. The subleading singularity in (3) is a simple pole and we can use our
theorem above. We choose the contour below:
5
C4
C3
r
C1 C2
if f (z) has no singularities along the positive real axis. Let us consider the integral
I
f (z) log zdz (5)
C
where we choose the contour C to be the one shown in figure. Such contours are often called keyhole
contours. The integrals along C2 and C4 are
Z Z 0
f ( x ) log xdx + f ( x ) (log x + 2i ) dx
0
Z
= 2i f ( x )dx.
0
6
C1
C4
C3 r
C2
This implies
Z
f ( x )dx = Res of f (z) log z inside C.
0 i
As an example consider
Z
1
dx.
0 x3 + 1
As we saw in class the cube roots of 1 are at
ei/3 , ei and ei5/3 .
On the other hand we have on the unit circle log z = log ei = i = Arg(z). Thus the resiudes of the
function
log z
z3 + 1
on the unit circle will be iArg(z) divided by the derivative of the denominatior:
iArg(z)
Res(z) = .
3z2
You can verify that the three residues at