Math185f09 Hw10sol
Math185f09 Hw10sol
FALL 2009/10
PROBLEM SET 10 SOLUTIONS
1. (a) Show that if f has a pole or an essential singularity at a, then e
f
has an essential singularity
at a.
Solution. If f has a pole of order m at a, then there exists > 0 and g : D(a, ) C
analytic, g(a) = 0, such that
f(z) =
g(z)
(z a)
m
for all z D
n=0
a
n
(z a)
n
.
Let
h(z) :=
n=0
a
n+m
(z a)
n
.
Then
1
f(z) = h(z) +
m1
n=0
a
n
(z a)
mn
and so
e
f(z)
= e
h(z)
m1
n=0
e
an
(za)
mn
=: e
h(z)
F(z).
Note that e
h(z)
is analytic and non-zero. If e
f(z)
has a pole or a removable singularity at a,
then F(z) = e
f(z)
e
h(z)
will have a pole or a removable singularity at a. So since F(z) has
an essential singularity at a, e
f(z)
must have an essential singularity at a.
If f has an essential singularity at a, then f(D
(a, )) = C (here S denotes the closure of the set S). Recall that if g is any continuous
function, then g(S) g(S). Since exp : C C is a continuous function, let S = f(D
(a, ))
and we have
exp(S) exp(S) = exp(C) = C
.
Hence, (exp f)(D
(0, 1) be
f(z) =
n=
a
n
z
n
.
Then by the integral formula for Laurent coecients,
a
n
=
1
2i
_
r
f(z)
z
n+1
dz
where
r
is given by z : [0, 2] C, z(t) = re
it
, and 0 < r < 1. Hence
|a
n
|
1
2
_
2
0
|f(re
it
)ire
it
|
|re
it
|
n+1
dt
=
1
2
_
2
0
|f(re
it
)|
r
n
dt
1
r
n
log
1
|re
it
|
1
2
_
2
0
dt
=
1
r
n
log
1
r
.
Since this true for all 0 < r < 1, for n = 1, 2, 3, . . . , we may take limit as r 0
+
to
get
|a
n
| lim
r0
+
r
n
log
1
r
= 0
(note that this only works when n is positive). Hence a
n
= 0 for all n < 0 and so
f(z) =
n=0
a
n
z
n
must have a removable singularity at 0. Upon dening f(0) = a
0
, we may assume that
f : D(0, 1) C is an analytic function. Now we may apply maximum modulus theorem to
see that for all 0 < r < 1,
max
|z|r
|f(z)| = max
|z|=r
|f(z)| max
|z|=r
log
1
|z|
= log
1
r
.
Hence
max
|z|<1
|f(z)| lim
r1
log
1
r
= 0,
2
and so f 0 on D(0, 1).
(b) Let f : C
C be analytic on C
and R,
f(z) = z +
1
z
+
for all z C
.
Solution. Since f has a pole of order 1 at 0 and is analytic otherwise, the Laurent
expansion of f takes the form
f(z) =
n=1
a
n
z
n
. (2.1)
Let be the closed curve z : [0, 2] C, z() = e
i
. Since f(e
i
) R for all [0, 2], we
must have
f(e
i
) = f(e
i
).
The integral formula for Laurent coecients yields, for any n Z,
a
n
=
1
2i
_
f(z)
z
n+1
dz
=
1
2
_
2
0
f(e
i
)e
in
dz
=
1
2
_
2
0
f(e
i
)e
in
dz
=
1
2
_
2
0
f(e
i
)e
in
dz.
Note that
a
n
=
1
2
_
2
0
f(e
i
)e
in
dz
and so we get
a
n
= a
n
for all n Z. By (2.1), a
n
= 0 for all n > 1. So a
n
= 0 for all n > 1. Let a
1
= and so
a
1
= a
1
= . Let a
0
= . Then (2.1) becomes
f(z) = z +
1
z
+ .
We know that = 0 since f has a simple pole at 0. We also know that R since
= a
0
= a
0
= .
3. Evaluate the integral
_
i
f
i
for i = a, b.
(a) f
a
: C
C is given by
f
a
(z) = e
e
1
z
and
a
is the boundary D(0, 2) traversed once counter-clockwise.
3
Solution. f
a
is analytic in C
(0, ) C is given by
f
b
(z) =
1
(sin z)
3
and
b
is the boundary D(0, 1) traversed once counter-clockwise.
Solution. f
b
is analytic in D
1
5!
z
4
+
__
3
=
1
z
3
_
1 + 3
_
1
3!
z
2
1
5!
z
4
+
_
+ 6
_
1
3!
z
2
1
5!
z
4
+
_
2
+
_
.
Now observe that the terms enclosed in the second parentheses onwards would all have
powers at least 4 and so will not contribute to the z
1
term. The only term in the rst
parentheses that contribute to the z
1
term is the z
2
term, which has coecient 3/3! = 1/2.
By the residue theorem,
_
b
f
b
= 2i Res(f
b
; 0) Ind(
b
; 0) = i.
4. (a) Let the Laurent expansion of cot(z) on A(0; 1, 2) be
cot(z) =
n=
a
n
z
n
.
Compute a
n
for n < 0.
Solution. Let be the circle D(0, r) traversed once counter-clockwise and 1 < r < 2.
Note that Ind(; z) = 1 for all z D(0, r), ie. the bounded component of . By the integral
formula for Laurent coecients,
a
k
=
1
2i
_
cot(z)
z
k+1
dz =
1
2i
_
z
k1
cot(z) dz
4
for k N. For k = 1,
z
k1
cot(z) = cot(z) =
cos(z)
sin(z)
has three isolated (non-removable) singularities in the bounded component of , namely,
1, 0, 1. So by the residue theorem
2
,
a
1
=
1
2i
_
cot(z) dz
= Res(cot(z); 1) + Res(cot(z); 0) + Res(cot(z); 1)
=
cos(z)
cos(z)
z=1
+
cos(z)
cos(z)
z=0
+
cos(z)
cos(z)
z=1
=
3
.
For k 2,
lim
z0
z[z
k1
cot(z)] = lim
z0
z
sin(z)
[z
k1
cos(z)]
=
1
_
lim
z0
z
sin(z)
_
_
lim
z0
z
k1
cos(z)
_
= 0
In other words, 0 is a removable singularity of z
k1
cot(z) for k 2. Note that the residue
about any removable singularity is 0. So by the residue theorem
1
,
a
k
=
1
2i
_
z
k1
cot(z) dz
= Res(z
k1
cot(z); 1) + Res(z
k1
cot(z); 0) + Res(z
k1
cot(z); 1)
=
z
k1
cos(z)
d
dz
sin(z)
z=1
+ 0 +
z
k1
cos(z)
d
dz
cos(z)
z=1
=
(1)
k1
+
1
=
_
_
_
0 if k is odd,
2
if k is even,
for k 2.
(b) For n = 0, 1, 2, . . . , compute
1
2i
_
n
dz
z
3
sin z
where
n
is the circle D(0, r
n
) traversed once counter-clockwise and r
n
= (n +
1
2
).
Solution. Note that for m = 0,
lim
zm
(z m)
1
z
3
sin z
= lim
zm
z m
z
3
sin(z m)
=
1
m
3
3
lim
zm
z m
sin(z m)
=
1
m
3
3
2
We use result that Res(/; a) = (a)/
(a) = 0.
5
while for m = 0
lim
z0
z
4
1
z
3
sin z
= lim
z0
z
sin z
= 1.
So the integrand z
3
csc z has a pole of order 3 at 0 and simple poles at m for all m Z,
m = 0. Since D(0, r
n
), the bounded component of
n
, contains {m | m = n, ,
1, 0, 1, . . . , n}, the residue theorem yields
1
2i
_
n
dz
z
3
sin z
=
n
m=n
Res
_
1
z
3
sin z
; m
_
.
Now for m = 0,
Res
_
1
z
3
sin z
; m
_
=
1/z
3
d
dz
sin z
z=m
=
1/m
3
3
cos(m)
=
(1)
m
m
3
3
,
and for m = 0,
Res
_
1
z
3
sin z
; 0
_
= 0.
For the latter, observe that the Laurent expansion of z
3
csc z contains only even powers,
and so a
1
= 0. Hence
1
2i
_
n
dz
z
3
sin z
=
n
m=1
(1)
m
m
3
3
n
m=1
(1)
m
m
3
3
= 0.
5. (a) Does the following function have an antiderivative on A(0; 4, )?
z
(z 1)(z 2)(z 3)
Solution. Let f : A(0; 4, ) C be
f(z) :=
z
(z 1)(z 2)(z 3)
.
Let A(0; 4, ) be the boundary of a rectangle R traversed once counter-clockwise.
Since A(0; 4, ), we must have 1, 2, 3 R, the bounded component of . So the
winding numbers Ind(; n) = 1 for n = 1, 2, 3. By the residue theorem
1
2i
_
f(z) dz = 0.
Note that this holds for arbitrary and thus all A(0; 4, ), = R for some R. Hence
by Problem Set 9, Problem 3(a) (which is really Moreras Theorem), f has an antiderivative
on A(0; 4, ).
6
(b) Does the following function have an antiderivative on A(0; 4, )?
z
2
(z 1)(z 2)(z 3)
Solution. Let g : A(0; 4, ) C be
g(z) :=
z
2
(z 1)(z 2)(z 3)
.
The same argument above applies but this time
Res(g; 1) = lim
z1
(z 1)g(z) = lim
z1
z
2
(z 2)(z 3)
=
1
2
,
Res(g; 2) = lim
z2
(z 2)g(z) = lim
z2
z
2
(z 1)(z 3)
= 4,
Res(g; 3) = lim
z3
(z 3)g(z) = lim
z3
z
2
(z 1)(z 2)
=
9
2
.
So
1
2i
_
g(z) dz = 1 = 0.
Recall that if g has an antiderivative on A(0; 4, ), then the integral about any closed curve
in A(0; 4, ) is necessarily
3
0. Hence g does not have an antiderivative on A(0; 4, ).
3
This doesnt depend on having a simply connected region. So it works for annulus too.
7