Problem 1.

Prove that Z
sin x
dx =
0 x 2
R R + RR
Proof. We note by the symmetry of f (x) = x1 sin x, 2 0
sin x
x =
sin x
x . We also know that R
sin x
x dx =
R R ix
Im R eix , and so we only have to calculate

eix eix
Z Z
lim dx + dx (1)
0,R l x l+ x

where we are using the contours shown below. By Cauchys theorem, we can conclude that

eix eiz ez
Z Z Z
dx = dz + dz (2)
x c z cR z

cR

c

l l+

Figure 1: Contours used in solution

Now note that on cR we have z = R exp (it) = R cos t + iR sin t, thus

Z Z
exp iz exp (iR cos t) exp (R sin t)
dz = iR exp (it)dt
CR z 0 R exp (it)

Then we obtain the estimate

exp iz  
Z  Z  Z Z /2
 
 dz  =  exp (iR cos t) exp (R sin t)i dt exp (R sin t) = 2 exp (R sin t)dt (3)

CR z 0 0 0

This integral on the right vanishes, because sin t 2t/ in the interval [0, /2] (proof: sin is concave), and
thus Z /2 Z /2     
2Rt R
exp (R sin t)dt exp dt = 1 exp
0 0 2R
Z
exp iz
and this clearly vanishes as R . thus 0 as R . Now it remains to evaluate
CR z
Z Z
exp iz 1
dz = + f (z)dz (4)
C z C z

1
Where f is bounded (say, by M ) as z 0. Then
Z Z
1 1
+ f (z)dz dz + M  (5)
C z C z

Thus C z1 + f (z)dz C z1 dz as  0. This second integral is easy to evaluate, and is i. Recalling that
R R
R
sin x/x = Im 21 eix /x dx, we obtain the desired result
R
0

Z
sin x
dx =
0 x 2

Problem 2. Prove that Z Z

sin x2 dx = cos x2 dx =
0 0 2 2
R 2
Proof. Here we will use the fact that ez dz =
, which has a standard proof (using polar coordinates)

R 2
which we will not reproduce here. By symmetry we can conclude that 0 ez dz = /2. We will now
2
integrate ez over the following contours:

c b 4

0 a
R

Figure 2: contours used in problem 2

Then we have Z
z 2
lim e dz =
R a 2
2
ez dz = 0. To see why this is so, we let z = R exp it, t [0, /4], so that
R
We also have that limR b
z 2 = R2 exp 2it = R2 cos 2t + iR2 sin 2t, and thus
Z  Z /4

 Z /4
2
 ez dz  =  2 2
exp R2 cos 2t dt
 



exp R cos 2t exp iR sin 2t iR exp (it)dt R

  
b 0  0

Since cos(2t) 1 4t/, this integral on the right is less than

/4




exp R2 1 1 exp R2
Z
exp R2 exp 4R2 t dt = R exp R2



R
0 4R2 R

and this clearly vanishes as R . Finally, we evaluate along c, where we have z = t exp (/4) t [0, ],

2
which gives us
Z Z Z
1+i
exp z 2 dz = exp it2 exp (/4)dt = cos t2 i sin t2



dt (6)
c 0 0 2

Performing more algebraic manipulations, we obtain

Z Z
1 2
cos t2 + sin t2 + i cos t2 sin t2 dt




exp z dz =
c 2 0

2 2
ez dz = ez =
R R
By cauchys theorem, we can conclude that c a
/2 + 0i, and thus

Z Z Z

2 2 2 2
cos t dt = sin t dt and cos t dt =
4
|0 {z 0
} 0
imaginary part

which is what we wanted to show.

Problem 3. Evaluate the integrals
Z Z
exp (ax) sin bx dx and exp (ax) cos bx dx
0 0

where a and b are positive real constants.

Solution: These are clearly the real and imaginary part of the integral
Z
exp ((a + ib)z)dz
0

to solve this problem, we will use the following contours

0
R

Z
Here the curve is the ray aligned with a + ib. First it is clear that lim exp ((a + ib)z)dz is the desired
R
integral. We also have
Z Z Z
a + ib
exp (a2 + b2 )t = 2


lim exp ((a + ib)z)dz = exp ((a + ib)(a + ib)t)(a+ib) = (a+ib)
R 0 0 a + b2
Z
It remains to show that exp ((a + ib)z)dz vanishes as R goes to infinity. To prove this, let be

parametrized by R exp (it), where t [0, ], where < /2; this follows from the fact that a, b are pos-

3
 itive. Then we have
Z Z
exp ((a + ib)z)dz = exp (R(a cos t + b sin t) + iR(b cos t a sin t))Ri exp (it)dt
0

Taking absolute values, we can conclude that this integral vanishes if the integral
Z /2
R exp (R(a cos t + b sin t))dt
0

vanishes. This is true because the minimum value of a cos t + b sin t in (0, /2), lets call it m, is clearly
positive and so since R exp (Rm) 0 as R , we have the desired result. Thus
Z Z
b a
exp (ax) sin bx dx = and exp (ax) cos bx dx = (7)
0 a2 + b2 0 a2 + b2

Problem 4 (Schwarz reflection principle). Let H be an open set and let = { z : Im z = 0 } be a

nonempty open subset of the real axis. Suppose that f is holomorphic in , continuous on , and takes
= , where here denotes the image of under the mapping z 7 z. Define
real values on . Let
by
the function F C()
f (z)
z
F (z) =
z

f (z)

Prove that F is holomorphic in .

Proof. The only thing to prove is that F is differentiable at every point z in its domain. We will prove this
in cases

Case 1. z , then by the openness of there exists a neighborhood N around z such that F = f on N . Since
f is holomorphic on N , F is holomorphic on N .

Case 2. z . It is clear that z 7 z is an open mapping, and so there exists a neighborhood N which contains
z and thus N contains z. Pick h so small so that z + h N , and consider the difference
 
F (z + h) F (z) f (z + h) f (z)
=
h h

By the holomorphicity of f in , we can conclude that the right side converges to f 0 (z) as h 0, that is

 
F (z + h) F (z) f (z + h) f (z)
lim = lim = f 0 (z)
h0 h h0 h

This proves that F is holomorphic in . Furthermore, since f (z) R for z , we can conclude that for
u we have limuz F (u) = limuz f (u) = f (z), and so F extends continuously to from both and .

Case 3. Now consider the case where z . Now I claim that for every point z in there exists a ball B which

4
is centered on z such that B H . To prove this claim, note that R and bdry are closed, and so
bdry R is closed. But since is also open by hypothesis, we must conclude that = R or = ,
since R is connected. However, this seems a bit too restrictive, and so we wont use it in our proof. Rather,
we shall prove a weaker claim that F is holomorphic on the interior of with respect to R, which we shall
denote by o (e.g. if = [0, 1], then o is (0, 1), not ). Then o is an open subset of bdry (with the
relative topology), and so for each point z of o there exists an open ball B around z which contains none
of bdry o . Now consider B H. By the fact that z , we can conclude that B H intersects , but
if B H also intersects the exterior of , then we can write

B H = (B H) exterior (B H)

Thus writing B H as the union of two non-empty disjoint open sets, contradicting the fact that B H is
connected. This proves that B H . With all that, we can conclude that B H = B H and

B R o , and so B is completely contained in .

Now we will show that F is holomorphic at z (which is the centre of B). To do, we will use Moreras theorem,
and prove that the integral of F over every triangle T in B vanishes. The proof is quite simple: if T does

not interesect the real axis, then we are done, for then the triangle is contained entirely in either or ,
where F is holomorphic. If the triangle does intersect the real axis, we can decompose the triangle as in the
following picture

Figure 3: Triangle intersecting R

Since this is the integral of F over two polygonal paths contained in the closure of and respectively,
we can conclude by continuity that the integral
which are the limits of triangles strictly contained in and ,
of F vanishes over the entire triangle. This proves that F is holomorphic on B.

Problem 5. Show that an entire analytic function with bounded real part must be constant.

Proof. We will use the fact that if f is entire and f is bounded then f is constant. This was proved in class,
and is known as Liouvilles theorem.

Suppose that f has bounded real part, then consider exp f . By the fact that |exp(f (z))| = exp(Re(f (z))),

5
we can conclude that that exp f is totally bounded. Thus exp f is constant, and so clearly f must also be
constant.
p
Problem 6. Let f be entire and suppose that |f (z)| M (1 + |z|) for all z C, with some constant
M > 0. Show that f is constant.

Proof. We will use that fact that if f is entire then

Z
k! f ()
f (k) (z) = d
2i CR ( z)k+1

for any circle CR (with radius R) containing z. Then we can conclude that
Z 

 
 (k)  k!  f () 
 d M (1 + R) k!
f (z) 
2 CR ( z)k+1
  Rk

Sending R , we see that f (k) (z) vanishes for all z and for all k 1. In particular f 0 (z) = 0 for all z and
thus f is constant on C.

Problem 7. Let f be an entire function satisfying |f (z)| as |z| . Show that f : C C is

surjective. Derive the fundamental theorem of algebra as a corollary.

Proof. First note that if f satisfies the hypotheses of the theorem, then so does f for any C. Then
1
there exists a disk D so large that |f | 1 outside this disk, and in particular, 1 outside of
f
D. Now since f is clearly not constant, we can conclude that 1/(f ) is also not constant, and thus
1/(f ) is unbounded. By the fact that 1/(f ) 1 outside of D, we must conclude that 1/(f ) is
unbounded on the inside of D. But this implies that 1/(f ) has some discontinuities in D, which can only
occur if there exists some z such that f (z) = . This proves surjectivity.

Now to prove the fundamental theorem of algebra as corollary, let f be a polynomial of degree n. Then
n n1
|f (z)| a |z| b |z| for some a, b R (this follows from f (z) = an z n + + a0 , and the triangle
n n1
inequality). But for large |z|, a |z| b |z| clearly gets arbitarily large, and thus |f (z)| as |z| .
We can therefore use the above theorem to prove that there exists a z such that f (z) = 0.