Review Basic Complex Analysis Complete

Math Math 213a (Fall 2021) Yum-Tong Siu 1
Review of Basic Complex Analysis
Complex Derivative and Cauchy-Riemann Equations. For a complex-

valued function f (z) of a complex variable defined in some open neighborhood
of c, the complex derivative f 0 (c) at c is defined as the limit of the difference
quotient
f (z + c) − f (c)
lim
z→c z−c
in the sense that given any ε > 0 there exists some δ > 0 such that

f (z + c) − f (c) 0

− f (c) < ε for 0 < |z − c| < δ
z−c
df
(if such a limit exists). We also denote f 0 (c) by dz
(c). A necessary condition
for the existence of f 0 (c) is that
∂f 1 ∂f
= = f 0 (c)
∂x i ∂y
at c, because restriction of the difference quotient to the horizontal line y = b
and to the vertical line x = a when passing to limit would give the same
result, where c = a + ib with a, b ∈ R. The reason for the factor 1i is that
z − c = i(y − b) with the factor i on x = a whereas z − c = x − a without
the factor i on y = b. This necessary condition
∂f 1 ∂f
=
∂x i ∂y
is known as the Cauchy-Riemann equations. When f (z) = u(x, y) + iv(x, y),
separation of the Cauchy-Riemann equations into real and imaginary parts
yields its alternative formulation
 ∂u ∂v
 ∂x = ∂y
 ∂u ∂v
∂y
= − ∂x .
Cauchy-Riemann Equations not Sufficient Condition for Existence of

Complex Derivative, without Assuming Some Smoothness Condition for
the Real and Imaginary Parts as Functions of Two Real Variables. The
Cauchy-Riemann equations for f (z) = u(x, y) + iv(x, y) at a point c = a + bi

are not sufficient for the existence of the complex derivative of f at z = c,
because they are satisfied at z = 0 if u(x, 0), u(0, y) and v(x, y) are all
identically zero, even without u(x, y) being continuous at 0, for example,
u(x, y) = x2xy
+y 2
for (x, y) 6= (0, 0) and u(0, 0) = 0.
Necessary and Sufficient Condition for Existence of Complex Derivative

Involving Cauchy-Riemann Equations. The complex derivative of f (z) =
u(x, y) + iv(x, y) exists at c = a + bi if and only if
∂u ∂v ∂u ∂v
(i) the Cauchy-Riemann equations ∂x
= ∂y
and ∂y
= − ∂x are satisfied at
(x, y) = (a, b), and
(ii) both u(x, y) and v(x, y), as functions of two real variables x, y, are
differentiable at (x, y) = (a, b) in the sense that

u(x, y) = u(a, b) + ux (a, b)(x − a) + uy (a, b)(y − a) + E1 (x, y)
v(x, y) = v(a, b) + vx (a, b)(x − a) + vy (a, b)(y − a) + E2 (x, y)


with
Ej (x, y)
lim p = 0 for j = 1, 2.
(x,y)→(a,b) (x − a)2 + (y − b)2
Proof. The existence of the complex derivative of f (z) = u(x, y) + iv(x, y)

at c = a + bi means that
f (z) = f (c) + f 0 (c)(z − c) + E(z)
with
E(z)
lim = 0 for j = 1, 2.
z→0 |z − c|
Taking the real and imaginary parts of

f (z) = f (c) + f 0 (c)(z − c) + E(z)
yield

u(x, y) = u(a, b) + ux (a, b)(x − a) + uy (a, b)(y − a) + Re E(z)
v(x, y) = v(a, b) + vx (a, b)(x − a) + vy (a, b)(y − a) + Im E(z).


On the other hand, when


(∗)

and  ∂u ∂v
 ∂x = ∂y
 ∂u ∂v
∂y
= − ∂x ,
we can add i times the second equation of (∗) to the first to get
u(x, y) + iv(x, y) = (u(a, b) + iv(a, b))

+ (ux (a, b) + ivx (a, b)) ((x − a) + i(y − b)) + (E1 (x, y) + iE2 (x, y),
which means
f (z) = f (c) + f 0 (c)(z − c) + (E1 + E2 )
when we set f 0 (c) = ux (a, b) + ivx (a, b).
Geometric Interpretation of Cauchy-Riemann Equations. The condition




with
Ej (x, y)
lim p = 0 for j = 1, 2
(x,y)→(a,b) (x − a)2 + (y − b)2
means that the map Φ : (x, y) 7→ (u, v) can be approximated at (x, y) = (a, b),
to an order higher than the first, by the linear map dΦ : R2 → R2 defined by
the 2 × 2 Jacobian matrix

ux (a, b) uy (a, b)
vx (a, b) vy (a, b)
which is the tangent map (or derivative) of Φ at (x, y) = (a, b).

When the R-vector space R2 is identified with C-vector space C, the

Cauchy-Riemann equations ux = vy and uy = −vx at (x, y) = (a, b) is
equivalent to the condition that the R-linear map dΦ at (a, b) from R2 to
R2 at (a, b) is C-linear as a map from C to C, because the operation in C of
multiplication by i is represented by the 2 × 2 matrix

0 −1
1 0
as an R-linear self-map of R2 and the Cauchy-Riemann equations ux = vy
and uy = −vx are equivalent to

0 −1 ux uy ux uy 0 −1
= .
1 0 vx vy vx vy 1 0
On the other hand, when we consider the complex-valued function f (z) of a
complex variable z, instead of the pair of real-valued functions u(x, y), v(x, y)
of two real variables x, y, the condition
f (z) = f (c) + f 0 (c)(z − c) + E(z)
with
E(z)
lim =0
z→c |z − c|
means that the map f can be approximated at z = c, to an order higher than

the first, by the C-linear map df : C → C defined by multiplication by f 0 (c).
The C-linear map df : C → C is the same as the R-linear map dΦ : R2 → R2
when the C-vector space C is identified with the R-vector space R2 .
Cauchy’s Theorem. Suppose the complex derivative of f (z) = u(x, y) +
iv(x, y) exists at every point of an open neighborhood Ω of a closed rectangle
R in C = R2 . If both u(x, y) and v(x, y) have continuous first-order partial
derivatives at every point of Ω, then
Z
f (z)dz
∂R
(where ∂R means the boundary of R in the counterclockwise sense) which is

defined as
Z Z Z
(u(x, y) + iv(x, y))d(x + iy) = (udx − vdy) + i (vdx + udy)
∂R ∂R ∂R
vanishes.
Proof. The theorem of Cauchy simply follows from Stokes’s theorem

Z Z
(udx − vdy) = (−vx + uy )dxdy = 0,
Z∂R ZR
(vdx + udy) = (ux − vy )dxdy = 0.
∂R R
The key additional assumption here is the existence and continuity of ux ,

uy , vx , vy at every point of Ω. The removal of this unnecessary additional
assumption is the contribution of Goursat and is done in the following the-
orem of Cauchy-Goursat, which contains one technique of applying Stokes’s
theorem without the condition of continuous derivative required by the fun-
damental theorem of calculus for Riemann integration.
Theorem of Cauchy-Goursat for a Rectangle. If the complex derivative

of f (z) = u(x, y) + iv(x, y) exists at every point of an open neighborhood Ω
of a closed rectangle R in C = R2 , then
Z
f (z)dz = 0.
∂R
Before giving a rigorous argument, let us first intuitively understand the

technique used. Though the real and imaginary parts of f are not assumed
continuously differentiable at every point of Ω, the existence of f 0 (c) for c ∈ Ω
means that f (z) can be approximated at c, with an error Ec (z) of an order
higher than the first, by a polynomial Pc (z) = f (c) + f 0 (c)(z − c) of degree
≤ 1 in the complex variable z.
For a small closed rectangle Rc containing c of total boundary length `c ,

we have
Z Z Z Z
f (z)dz = Pc (z)dz + Ec (z)dz = Ec (z)dz,
∂Rc ∂Rc ∂Rc ∂Rc
by Cauchy’s theorem applied to Pc (z), with

Z
Ec (z)dz = o `c 2

∂Rc
(where o (`c ) is the Landau symbol for lower order), in which one order of `c
is from the length of ∂Rc and the other from Ec (z). We divide up each side
of R into m equal parts so that we end up with m2 equal closed subrectangles
Rcj with a point cj inside (for 1 ≤ j ≤ m2 ).
From the above argument applied to each Rcj to get

Z
1
f (z)dz = o

m2

∂Rcj
and
m2 Z
Z
X
2 1
f (z)dz ≤ f (z)dz = m o = o(1),

m2

∂R j=1
∂Rcj
R
which implies, as m → ∞, the vanishing of ∂R f (z)dz. This argument is
not yet rigorous, because there is a question of uniformity in the use of the
Landau symbol of lower order. The usual argument for uniformity by com-
pactness is by reductio ad absurdum of selecting a sequence of points where
the uniformity fails and considering the situation at one of their accumula-
tion points. We now use the following rigorous argument of combining the
above intuitive analysis with the uniformity argument by compactness.
R
Suppose ∂R f (z)dz is nonzero. Let A be its absolute value. Let ` be the
length of the boundary ∂R of R. Divide R into 4 equal closed subrectangles.
Then one closed subrectangle R(1) out of the four must satisfy
Z
A
f (z)dz ≥ .

∂R (1)
4
Divide up R(1) into closed subrectangles. Then one closed subrectangle R(2)
out of the four must satisfy
Z
≥ A.

f (z)dz 42
∂R(2)
Continue this process to end up with a sequence of closed subrectangles R(m)

of total boundary length 2`m such that
Z
≥ A.

f (z)dz 4m
∂R(m)
Now use compactness of R to conclude that there is some c ∈ R which is in

every R(m) in the nested sequence of closed subrectangles. Existence of f 0 (c)
means that f (z) = Pc (z) + Ec (z) with
Z
Pc (z)dz = 0 for m ∈ N
∂R(m)
and the estimate that for any ε > 0 there exists mε ∈ N such that |Ec (z)| ≤
ε 2`m on R(m) and
Z 2
`
Ec (z)dz ≤ ε m
for m ≥ mε .
(m)
∂R 2
This yields the contradiction
2 Z
` A
ε m ≥
f (z)dz ≥ m
2 ∂R(m) 4
when ε`2 < A and m ≥ mε .
Theorem of Cauchy-Goursat (for Domain with Piecewise Smooth Bound-

ary). We use the terminology that a function f (z) is holomorphic or complex-
analytic on an open subset Ω of C if the complex derivative of f (z) exists
at every point of Ω. The theorem of Cauchy-Goursat for a closed rectangle
implies the existence of a primitive F (z) for a holomorphic function f (z) on
a simply connected domain Ω in C in the sense that the complex derivative
F 0 (z) is equal to f (z) at every point of Ω. The primitive F (z) is constructed
by fixing a point z0 and define
Z
F (z) = f (z)dz
Cz
where Cz is any broken path joining z0 to z with each line-segment either

horizontal or vertical.
The theorem of Cauchy-Goursat for a closed rectangle guarantees that

F (z) is independent of the choice of Cz . By applying the fundamental the-
orem of calculus to a broken path whose last line-segment is horizontal, we
conclude that ∂F∂x
= f . By applying the fundamental theorem of calculus to
a broken path whose last line-segment is vertical, we conclude that ∂F∂y
= i f.
Thus, the Cauchy-Riemann equations for F are satisfied at every point of Ω.
Since both first-order partial derivatives Fx and Fy of F are continuous on Ω,
both the real part and the imaginary part of F are differentiable functions
of x, y at every point of Ω, by a standard application of the mean-value the-
orem. Thus the complex derivative F 0 of F and is equal to f at every point
of Ω. Once we have a primitive F of f on Ω we have the following theorem
of Cauchy-Goursat for a general piecewise smooth contour.
Theorem of Cauchy-Goursat. If f (z) is a holomorphic function on an open

subset Ω of C and G is a relatively compact open subset of Ω with piecewise
smooth boundary, then Z
f (z)dz = 0.
∂G
Piecewise smooth boundary property of G means that the boundary of G is
composed of a finite number of closed curve segments represented by continu-
ously differentiable maps from a closed interval of R into Ω with nowhere zero
derivatives, with two adjacent closed curve segments joined at end-points.
The reason for the piecewise smooth condition is that for a curve-segment
γ defined by t 7→ z = ϕ(t) ∈ Ω for t ∈ [a, b],
Z Z b
f (z)dz = f (ϕ(t))ϕ0 (t)dt
γ t=a
as a Riemann integral requires continuity of ϕ0 (t). The boundary of G is

allowed to have more than one component. The direction of travel along
boundary of G is the rule of ”driving on the right” (i.e., counter-clockwise)
so that the main part of the domain G is always on the left when one travels
along the boundary of G. The vanishing of
Z
f (z)dz = 0
∂G
comes from the application of the fundamental theorem of calculus which

yields Z
f (z)dz = F (c2 ) − F (c1 )
γ
for any closed curve-segment γ in Ω with initial point c1 and terminal point
c2 .
In the case of more than one component in the boundary of G, curve-

segments have to be inserted as cross-cuts in G joining different components
of the boundary so that the integration of f (z)dz over ∂G is equal to the
integration of f (z)dz over a single closed curve which includes traversing each
inserted cross-cut precisely twice in opposite directions.
Integral of Power Function along a Circle. For any ρ > 0 and any m ∈ Z,
Explicit computation with the parametrization θ 7→ ρeiθ (0 ≤ θ ≤ 2π) for
the circle |z| = ρ yields
Z Z 2π
m
z dz = i ρm+1 ei(m+1)θ dθ
|z|=ρ θ=0

i
(m+1)θ 2π
 m+1 e θ=0
= 0 for m 6= −1
=
2πi for m = −1

This explicit computation is used in the proof of Cauchy’s integral formula

and in the recovery of the coefficients of a power series or a Laurent series
from the integral of the series along a circle. The recovery of such coefficients
is used in establishing the uniqueness of power series and Laurent in series
in the disk or annulus of convergence.
Cauchy’s Integral Formula. If f (z) is a holomorphic function on an open

subset Ω of C and G is a relatively compact open subset of Ω with piecewise
smooth boundary, then
Z
1 f (z)
f (c) = dz for c ∈ G.
2πi ∂G z − c
Cauchy’s integral formula is established by choosing ρ > 0 small enough for
the closed disk Dρ0 (c) of radius ρ0 centered at c to be contained in G, and
applying the theorem of Cauchy-Goursat to fz−c (z)
on Ω − {c} to yield
Z
f (z)
dz = 0
∂ (G−Dρ (c)) z − c
for 0 < ρ ≤ ρ0 , which means that the integral

Z
f (z)
dz
∂G z − c
is equal to
f (z) − f (c)
Z Z Z
f (z) dz
dz = dz + f (c)
|z−c|=ρ z−c |z−c|=ρ z−c |z−c|=ρ z−c
= 2πi f (c)
by letting ρ → 0, because
Z
f (z) − f (c) f (z) − f (c)
dz ≤ 2πρ sup

|z−c|=ρ z − c |z−c|=ρ z − c
and
f (z) − f (c)
sup < ∞.
0<|z−c|≤ρ0 z−c
Differentiation of Definite Integral with respect to a Real Parameter of

Integrand. Suppose the complex-valued function f (x, t) is continuous on the
rectangle [a, b] × (c, d) in R2 such that ∂f
∂t
is also continuous on [a, b] × (c, d).
Then
d b
Z Z b
∂f
f (x, t)dx = (x, t)dx.
dt x=a x=a ∂t
This is the integral version of the statement that an infinite sum can be
differentiated term-by-term if the resulting infinite sum converges uniformly.
Both versions use the fundamental theorem of calculus in essentially the same
way.
For c0 ∈ (c, d), by Fubini’s theorem of exchanging the order of integration

and by the fundamental theorem of calculus
Z τ Z b Z b Z τ
∂f ∂f
(x, t)dx dt = (x, t)dt dx
t=c0 x=a ∂t x=a t=c0 ∂t
Z b Z b
= f (x, τ )dx − f (x, c0 )
x=a x=a
for τ ∈ (c, d). By differentiating with respect to τ and using the fundamental
theorem of calculus we get with τ = t
Z b
∂ b
Z
∂f
(x, t)dx = f (x, t)dx.
x=a ∂t ∂t x=a
Cauchy’s Integral Formula for Derivatives of Holomorphic Functions.

If f (z) is a holomorphic function on an open subset Ω of C and G is a
relatively compact open subset of Ω with piecewise smooth boundary, then
for any nonnegative integer n the n-th complex derivative f (n) (z) of f is given
by Z
(n) n! f (ζ)
f (z) = dζ for z ∈ G.
2πi ∂G (ζ − z)n+1
To prove it, we first verify that for any complex-valued continuous function
g on ∂G and for any nonnegative integer n, the function
Z
g(ζ)
z 7→ h(z) = n+1
dζ
∂G (ζ − z)
is infinitely differentiable as a function of x and y (where z = x + iy) and is

holomorphic on G.
The infinite differentiability of h follows from
∂ p+q ∂ p+q
Z
1
(†)p,q h(z) = g(ζ) p q dζ
∂xp ∂y q ∂G ∂x ∂y ζ −z
for p, q ∈ N ∪ {0}, which is obtained by applying the above result on differ-

entiating a definite integral repeatedly with respect to the real parameter x
1
or y of the integrand. Since ζ−z as a function of z (for fixed ζ) satisfies the
Cauchy-Riemann equations, from (†)p,q with (p, q) = (1, 0) and (p, q) = (1, 0),
we conclude that h(z) satisfies the Cauchy-Riemann equations and is there-
fore holomorphic on G. Note that when we apply the result on differentiating
a definite integral repeatedly with respect to a real parameter of the inte-
grand, we can only differentiate with respect to x or y and not directly with
respect to the complex parameter z though we can indirectly do it.
We are now ready to prove Cauchy’s integral formula for derivatives of

holomorphic function
Z
(n) n! f (ζ)
(\)n f (z) = dζ for z ∈ G
2πi ∂G (ζ − z)n+1
by induction on the nonnegative integer n. The case of n is simply the

Cauchy integral formula for holomorphic functions. We now go from the n-
th step to the (n + 1)-th step. By applying what we just proved to the case
of g(ζ) = f (ζ), we conclude from (\)n that f (n) is infinitely differentiable as
a function of x, y and is holomorphic on G. Differentiating both sides of (\)n

with respect to the real parameter x in the integrand, we obtain
∂ (n)
f (n+1) (z) = f (z)
∂x Z
n! ∂ 1
= f (ζ) dζ
2πi ∂G ∂x (ζ − z)n+1
Z
n! d 1
= f (ζ) dζ
2πi ∂G dz (ζ − z)n+1
Z
(n + 1)! f (ζ)
= n+2
dζ
2πi ∂G (ζ − z)
This concludes the proof of Cauchy’s integral formula for derivatives of holo-
morphic functions. -A corollary is that any holomorphic function of z = x+iy
is infinitely differentiable as a function of x, y.
Uniform Limit of Holomorphic Functions. From the Cauchy integral for-

mula for derivatives of holomorphic functions, it follows that if fn (z) is a
sequence of holomorphic functions on an open subset Ω of C such that fn (z)
converges uniformly on compact subsets to a function f (z) on Ω, then f (z)
(k)
is holomorphic on Ω and for any positive integer k the k-th derivative fn (z)
of fn (z) converges uniformly on compact subsets of Ω to the k-th deriva-
tive f (k) (z) of f (z). As a consequence the function on an open disk defined
by a convergent power series is holomorphic and the power series can be
differentiated term by term.
Cauchy’s Kernel. Cauchy’s integral formula

Z
1 f (ζ)
f (z) = dζ
2πi ∂G ζ −z
means that f (z) can be interpreted the limit of the Riemann sum
X 1 f (ζj )
(ζj+1 − ζj ),
j
2πi ζj − z
which means that any holomorphic function f (z) is the limit of a finite lin-
ear combination, with constant coefficients, of the translates of the special
function z1 , which is the Cauchy kernel.. The word “kernel” carries the mean-
ing that all holomorphic functions germinate from the Cauchy kernel . The
1 1
Cauchy kernel is alternatively taken to mean 2πi ζ−z
so that any holomorphic
function is reproduced by integration against it along a surrounding contour.
Relation Between Cauchy’s Kernel and Newton’s Potential. The Cauchy

kernel is related to Newton’s potential on the plane, which is described by
the equation
1
∆ log |z| = δ0
2π
in the sense of distributions, where δ0 is the Dirac delta at the origin. The
denominator 2π is the circumference of the unit circle, because by the diver-
gence theorem
Z Z Z
∆ log |z| = (∂~n log |z|) ds = ds = 2π,
D |z|=1 |z|=1
where ∂~n is the derivative in the normal direction of the boundary |z| = 1 of
the unit disk D. To facilitate the discussion of the relation between Cauchy’s
kernel and Newton’s potential, we now introduce two differential operators
∂
∂z
and ∂∂z̄ .
Consider the complex-valued differential dz = dx + idy and dz̄ = dx − idy

and their dual vector fields

∂ 1 ∂ ∂
= −i ,
∂z 2 ∂x ∂y

∂ 1 ∂ ∂
= +i
∂ z̄ 2 ∂x ∂y
Here a vector field is identified the differentiation along it so that the vector
∂
field with components (a, b) is identified with the differential operator a ∂x +
∂ ∂ ∂
b ∂y . dz and dz̄ are dual to ∂z and ∂ z̄ in the sense that the evaluation of the
∂
1-form dz is 1 at the vector field ∂z and is 0 at the vector field ∂∂z̄ whereas the
∂
evaluation of the 1-form dz̄ is 0 at the vector field ∂z and is 1 at the vector
∂
field ∂ z̄ .
The Cauchy-Riemann equations ∂f ∂x

= 1i ∂f
∂y
can be rewritten simply as
∂f 0
∂ z̄
= 0. The complex derivative f (z) of a holomorphic function f is equal
∂f
to ∂z
, because
∂f ∂f ∂f ∂f
f0 = = + = .
∂x ∂z ∂ z̄ ∂z
∂ ∂ 2 2 ∂ 2
The Laplacian ∆ = ∂x 2 + ∂y 2 is equal to 4 ∂z∂ z̄ . More precisely, the equation
for the Newton potential should be written as

1
∆ log |z| (dx ∧ dy) = δ0 ,
2π
because the right-hand side yields a number when evaluated at a function.
Since
i
dx ∧ dy = dz ∧ dz̄,
2
the equation for the Newton potential can be rewritten as
∂2

1 i
4 log |z| dz ∧ dz̄ = δ0 .
2π ∂z∂ z̄ 2
or
−1 ∂ 2

2
log |z| dz ∧ dz̄ = δ0 ,
2πi ∂z∂ z̄
which, after using
∂ ∂ ∂ 1
log |z|2 = (log z + log z̄) = log z = ,
∂z ∂z ∂z z
becomes
−1 ∂ 1
dz ∧ dz̄ = δ0
2πi ∂ z̄ z
1
where the Cauchy kernel z
appears. Note that
∂ 1
log z =
∂z z
holds in the sense of distributions, because the function z1 is locally inte-
grable on C so that integration by parts can be justified when both sides are
integrated against a smooth function on C with compact support.
Cauchy Integral Formula for Smooth Functions. After a translation and

a notational change for the variable, we can rewrite the above equation as

−1 ∂ 1
dζ ∧ dζ̄ = δz .
2πi ∂ ζ̄ ζ − z
Suppose G is a bounded domain in C with piecewise smooth boundary and

z ∈ G. Suppose f is a C ∞ function on an open neighborhood of the closure
Ḡ of G in C. Then exterior differentiation of 1-form (using the basis dz, dz̄
instead of dx, dy with z replaced by ζ) yields, in the sense of distribution,
∂f
(ζ)dζ̄ ∧ dζ

−1 f (ζ)dζ −1 ∂ ζ̄ −1 ∂ 1
d = + f (ζ) dζ̄ ∧ dζ
2πi ζ − z 2πi ζ −z 2πi ∂ ζ̄ ζ − z
∂f
−1 ∂ ζ̄
(ζ)dζ̄ ∧ dζ
= − f (ζ)δz .
2πi ζ −z
In the above computation, on the 1-form which contains dζ the exterior
differential operator d is simply equal to (dζ̄)⊗ ∂∂ζ̄ . Applying Stokes’s theorem
to integration over G yields
∂f
−1
Z
f (ζ)dζ −1
Z
∂ ζ̄
(ζ)dζ̄ ∧ dζ
= − f (z),
2πi ∂G ζ −z 2πi G ζ −z
or
∂f
1
Z
f (ζ)dζ 1
Z
∂ ζ̄
(ζ)dζ ∧ dζ̄
f (z) = + ,
which is Cauchy’s integral formula for smooth functions. Distributions are
1
used because of the singularity ζ−z . To avoid rigorous details involving dis-
tributions, one can simply argue by removing a small disk centered at z and
then pass to limit as the radius of the small disk tends to 0. We will leave
the derivation without distribution theory to a homework problem.
Cauchy’s Integral Formula for Smooth Functions as Complex-Analytic

Analogue of Fundamental Theorem of Calculus. The fundamental theo-
rem of calculus states that
Z x
dg(t)
g(x) = g(a) + dt,
t=a dt
which we can rewrite as

Z b
dg(t)
g(x) = g(a) + χ(−∞,x) (t) dt,
t=a dt
where χ(−∞,x) is the characteristic function of the interval (−∞, x). The
rewriting is motivated by the fact that the derivative of the Heaviside function
is the Dirac delta function, which is analogous to the equation for Cauchy
kernel derived from Newton’s potential.
Cauchy’s integral formula for smooth functions states that
∂f
1
Z
f (ζ)dζ 1
Z
∂ ζ̄
(ζ)dζ ∧ dζ̄
f (z) = + .
1
R f (ζ)dζ
The term 2πi ∂G ζ−z
is analogous to the term g(z). The operator ∂∂ζ̄ is anal-
ogous to dtd . The Cauchy kernel 2πi 1 1
ζ−z
is analogous to χ(−∞,x) . Analogous
to Z b
d
χ(−∞,x) (t)ϕ(t) dt = ϕ(x)
dx t=a
from the fundamental theorem of calculus, the Cauchy integral formula for
smooth functions can be used to yield
h(ζ)dζ ∧ dζ̄
Z
∂ 1
= h(z)
∂ z̄ 2πi G ζ −z
for z ∈ G when h is a smooth function on Ω. Suppose we know that there
exists some smooth function F on a neighborhood U of Ḡ such that
∂F
= h(z)
∂ z̄
on U . Then we can apply Cauchy’s integral formula for smooth functions to
F to get
Z Z ∂F (ζ)dζ ∧ dζ̄
1 F (ζ)dζ 1 ∂ ζ̄
F (z) = +
h(ζ)(ζ) dζ ∧ dζ̄
Z Z
1 F (ζ)dζ 1
= + ,
∂
which, upon applying ,
yields immediately
∂ z̄

h(ζ)(ζ) dζ ∧ dζ̄
Z
∂ 1
h(z) =
for z ∈ G. However, we still have the problem of coming up with some such
smooth solution F of
∂F
= h(z)
∂ z̄
on U .
The way to do it without using some other solution F (z) is to apply

Stokes’s theorem to the
∂h(ζ) dζ
d h(ζ) log |ζ − z|2 dζ̄ = log |ζ − z|2 dζ ∧ dζ̄ + h(ζ)
∂ζ ζ −z
on G (where z is fixed and ζ is the variable of C) to get
dζ ∧ dζ
Z Z Z
2 ∂h(ζ) 2
h(ζ) log |ζ − z| dζ̄ = log |ζ − z| dζ ∧ dζ̄ + h(ζ) .
∂G G ∂ζ G ζ −z
Finally we get the result
h(ζ)dζ ∧ dζ̄
Z
∂ 1
= h(z)
by applying ∂∂z̄ to both sides and using the Cauchy integral formula for
smooth functions for the complex-conjugate h̄ of h. We will leave the details
to a homework problem.
Power Series and Laurent Series Expansion of Cauchy Kernel. Since

1
holomorphic functions are generated by the Cauchy kernel ζ−z in integrals,
some properties of the Cauchy kernel pass over to holomorphic functions. One
such property is the power series expansion and Laurent series expansion of
the Cauchy kernel by using geometric series. For a ∈ C, when |z −a| < |ζ −a|
the geometric series expansion
X (z − a)n∞
1 1 1 1
= = =
ζ −z (ζ − a) − (z − a) ζ − a 1 − z−a
ζ−a n=0
(ζ − a)n+1
converges and gives a power series expansion of the Cauchy kernel.

For a ∈ C, when |z − a| > |ζ − a|, with the exchange of the roles of z and
ζ, the geometric series expansion
1 −1 −1 1
= = ζ−a
ζ −z (z − a) − (ζ − a) z − a 1 − z−a
∞ −1
X (ζ − a)n X (z − a)m
=− = −
n=0
(z − a)n+1 m=−∞
(ζ − a)m+1
(by using n = −(m + 1) in the change of dummy index)
converges and gives a Laurent series expansion of the Cauchy kernel
Power Series Expansion of Holomorphic Function on Disk. Suppose f is

a holomorphic function on |z − a| < R. For z in the disk |z − a| < R, choose
0 < r < R with |z − a| < r to get the power series expansion
∞
!
(z − a)n
Z Z
1 f (ζ)dζ 1 X
f (z) = = f (ζ) n+1
dζ
2πi |ζ|=r ζ − a 2πi |ζ|=r n=0
(ζ − a)
∞ ∞
f (n) (a)
Z
X 1 f (ζ) n
X
= n+1
dζ (z − a) = (z − a)n ,
n=0
2πi |ζ|=r (ζ − a) n=0
n!
where for the last step the Cauchy integral formula for derivatives of holo-
morphic functions is used.
Uniqueness of Power Series Expansion of Holomorphic Function on

Disk. The power series expansion of a holomorphic function f on an open
disk |z − a| < R is unique, because if f (z) is equal to the convergent power
series ∞
X
cn (z − a)n
n=0
on |z − a| < R, then for 0 < r < R and any nonnegative integer k,

Z ∞ Z
f (z)dz X
= cn (z − a)n−k−1 dz = 2πick
|z−a|=r (z − a)k+1 n=0 |z−a|=r
(n)
so that cn = f n!(a) by the Cauchy integral formula for derivatives of holo-
morphic functions.
Laurent Series Expansion of Holomorphic Function on Annulus. Sup-

pose f is a holomorphic function on R1 < |z − a| < R2 . For z in the annulus
disk R1 < |z − a| < R2 , choose R1 < r1 < r < r2 < R2 with r1 < |z − a| < r2
to get the Laurent series expansion
Z Z
1 f (ζ)dζ 1 f (ζ)dζ
f (z) = −
2πi |ζ|=r2 ζ − a 2πi |ζ|=r1 ζ − a
∞
!
(z − a)n
Z
1 X
= f (ζ) dζ
2πi |ζ|=r1 n=0
(ζ − a)n+1
−1
!
(z − a)n
Z
1 X
− f (ζ) dζ
2πi |ζ|=r1 n=−∞
(ζ − a)n+1
∞ Z
X 1 f (ζ)
= n+1
dζ (z − a)n ,
n=−∞
2πi |ζ|=r (ζ − a)
where for the last step Cauchy’s theorem is used.
Uniqueness of Laurent Series Expansion of Holomorphic Function on

Annulus. The Laurent series expansion of a holomorphic function f on an
open annulus R1 <P |z − a| < R2 is unique, because Suppose f (z) is equal to
the Laurent series ∞ a)n on R1 < |z − a| < R2 in the sense that
n=−∞ cn (z −P
both series n=0 cn (z − a)n and −1
P∞ n
n=−∞ cn (z − a) converge at every point
PnR1 < |z − a| k< R2 instead of just the convergence of the principal sum
of
k=−n ck (z − a) as n → ∞. Then for R1 < r < R2 and any integer k,
Z ∞ Z
f (z)dz X
= cn (z − a)n−k−1 dz = 2πick ,
|z−a|=r (z − a)k+1 n=0 |z−a|=r
yielding the uniqueness of ck .
Identity Theorem for Holomorphc Functions as Consequence of Power

Series Expansion on Open Disk.
Theorem. The zero-set of a non-identically-zero holomorphic function f on

a connected open subset Ω of C must be isolated.
As a corollary, any two holomorphic functions on a connected open subset

which agree on a set with an accumulation point must be identical.
To prove the theorem, let Z be zero-set of f in Ω. For a ∈ Z, we have a

power series expansion
∞
X
f (z) = cn (z − a)n
n=0
on any open disk |z − a| < r whose closure is contained in Ω. Either all the
coefficients of the power series vanish, in which case f is identically zero on
|z − a| < r and Z contains an open neighborhood of a or f (z) = (z − a)k g(z)
with ∞
X
g(z) = cn (z − a)n−k
n=k
nonzero at z = a for some nonnegative integer k, in which case a is an

isolated point of Z. This means that the set of all accumulation points of
Z, if nonempty, is both open and closed and therefore must be equal to the
connected set Ω, which contradicts the assumption that f is not identically
zero on Ω.
Isolated Singularities of Holomorphic Functions Classified by Principal

Part of Laurent Series. A point a of C is called an isolated singularity of a
holomorphic function f is f is holomorphic on some punctured neighborhood
0 < |z − a| < R of a. Let
∞
X
f (z) = cn (z − a)n
n=−∞
be the Laurent series expansion of f on the open annulus 0 < |z − a| < R

(with degenerate inner radius). The part
−1
X
cn (z − a)n
n=−∞
is called the principal part of the Laurent series. The isolated singularity of
f at a is called
• a removable singularity if the Laurent series has no principal part;
• a pole of order k if the principal part of

Pthe Laurent series consists of
−1
only a finite number of nonzero terms n=−k cn (z − a)n with c−k 6= 0;
• an essential singularity if the principal part of the Laurent series con-

tains an infinite number of nonzero terms.
A meromorphic function on an open subset G of C means a function f on

G − Z, where Z is a (possibly empty) discrete subset of G, such that every
point of Z is a pole of f .
Residue of Isolated Singularity of Holomorphic Function. The residue of

a holomorphic function f at an isolated singularity a, denoted by Resa (f ), is
defined as the coefficient c−1 of the first negative power of the Laurent series
expansion
X∞
f (z) = cn (z − a)n ,
n=−∞
which is also equal to Z

1
f (z)dz
2πi |z−a|=ε
for any sufficiently small positive number ε. If the isolated singularity of f

at a is a pole of order ≤ k (where k is any positive integer), the residue of f
at a can be alternatively computed by the formula
k−1
1 d k

Resa (f ) = (z − a) f (z) ,
(k − 1)! dz k−1 z=a
because (z − a)k f (z) is a power series whose (k − 1)-th power term has coef-
ficient c−1 . Residues are introduced for the computation of definite integrals
by techniques of complex analysis.
Characterization of Three Kinds of Isolated Singularities of Holomor-

phic Function. The isolated singularity of a holomorphic function f at a
point a is removable if and only if f is square integrable on some deleted
neighborhood of a, because, computed in polar coordinates (r, θ) centered at
a (i.e., z = a + reiθ ) with area element rdrdθ,

2
Z X ∞
n
c (z − a)

n
ε<|z−a|<ρ n=−∞
X∞ Z
= cm (z − a)m cn (z − a)n
m,n=−∞ ε<|z−a|<ρ
X∞ Z ρ Z 2π
m imθ n −inθ
= cm r e cn r e dθ rdr
m,n=−∞ r=ε θ=0
∞
X Z ρ
2
= 2π |cn | r2n+1 dr
n=−∞ r=ε
 
2n+2 2n+2
ρ X ρ −ε
= 2π |c−1 |2 log + |cn |2  ,
ε 2n + 2
n∈Z−{0}
which can remain bounded as ε → 0+ if and only if cn = 0 for all n < 0.
The isolated singularity of a holomorphic function f at a point a is a pole

of order k (for some positive integer k) if and only if |f (z)| is comparable to
1
|z−a|k
as z → a in the sense that
A B
≤ |f (z)| ≤
|z − a|k |z − a|k
for some positive constants A, B when z is in a sufficiently small deleted
1
neighborhood of a, because |f (z)| ∼ |z−a| k means that the isolated singularity
at a of the holomorphic function (z − a)k f (z), which uniformly bounded on

some deleted neighborhood of a, is removable and has a nonzero value at a.
The isolated singularity of a holomorphic function f at a point a is an

essential singularity if and only if the image of any deleted neighborhood
0 < |z − a| < r under f is dense in C, because such an image cannot be
dense in C in the case of removable or pole singularity and if one such image
1
is disjoint from some nonempty open disk |z − b| ≥ δ, then f (z)−b is uniformly
bounded on some deleted neighborhood of a and is holomorphic function on
some neighborhood of a, making it possible to conclude that the isolated
singularity of f (z) is either removable or a pole.
By the big Picard theorem (which follows from Nevanlinna’s theory of

defect relation to be discussed later), the image of a holomorphic function
near an essential singularity cannot miss more than two points.
Argument Principle and Rouché’s Theorem (Counting of Zeros and

Poles from Residues of Logarithmic Derivative). Let G be a bounded do-
main in C with piecewise smooth boundary and Ω be an open neighborhood
of the closure Ḡ of G. Let f be a meromorphic function on Ω without any
zeroes or poles on the boundary ∂G of G. Let a1 , · · · , ak be the set of all
zeroes of f in G with multiplicities p1 , · · · , pk and let b1 , · · · , b` be the set
of all poles of f in G with multiplicities q1 , · · · , q` . By using the residues
0
of the logarithmic derivative (log f )0 of f defined as ff , we can compute
Pk P`
µ=1 pµ − ν=1 qν as follows. Let
(z − a1 )p1 · · · (z − ak )pk
f (z) = g(z)
(z − b1 )q1 · · · (z − b` )q`
so that g(z) is holomorphic and nonzero at every point of Ḡ and

k `
f 0 X pµ X qν g0
= − + .
f µ=1
z − aµ ν=1 z − bν g
Then
k `
f0 X f0 X f0
Z
1
= Resaµ + Resbν
2πi ∂G f µ=1
f ν=1
f
k `
X pµ X qν
= Resaµ − Resbν
µ=1
z − aµ ν=1 z − bν
k
X `
X
= pµ − qν .
µ=1 ν=1
f0
Since f
= (log f )0 , it follows that
f0
Z Z
1 1
= d log f
2πi ∂G f 2πi ∂G
Z
1
= (d log |f | + i arg f )
2πi ∂G
Z
1
= d arg f
2π ∂G
1
= ∆∂G arg f.
2π
Here arg f is the argument of f which means θ when f = |f |eiθ . Though
arg f can be defined as an angle or direction, yet its numerical value is not
well-defined and can only be chosen up to an integral multiple of 2π.
We can choose a point P in a component γ of ∂G and arbitrarily choose

the value of the multi-valued function arg f at P and let it vary continuously
along the positive direction of γ until back at P again where the new value
may be different. The new value minus the old value of arg f is the increase
in the argument of f along γ, which we denote by ∆γ f . Its sum over all
components γ of ∂G is denoted by ∆∂G arg f , which we refer to as the total
increase of arg f along ∂G.
Argument Principle. For a meromorphic function f on an open neighbor-

hood of a bounded domain G with piecewise smooth boundary which does
not have any zero or pole on the boundary ∂G of G, the total number of
1
zeroes of f in G minus the total number of poles of f in G is equal to 2π
times the total increase of the argument of f along ∂G.
We now look at the effect of a “small” perturbation of f on the total

increase of the argument of f along ∂G. Suppose we have another meromor-
phic function g on Ω without zero or pole on ∂G such that |g| < |f | on ∂G.
Rouché’s theorem tells us that the total increase of the argument of f + g
along ∂G is the same as the total increase of the argument of f along ∂G so
that the total number of zeroes of f + g in G minus the total number of poles
of f + g in G is the same as the total number of zeroes of f in G minus the
total number of poles of f in G. The proof of Rouché’s theorem is as follows.
The function
f 0 + tg 0
Z
1 1
t 7→ = ∆∂G arg(f + tg)
2πi ∂G f + tg 2π
for 0 ≤ t ≤ 1 is continuous with integral values so that it is independent of
t ∈ [0, 1], because f (z) + tg(z) is nowhere zero for t ∈ [0, 1] and z ∈ ∂G from
the assumption that |f | > |g| on ∂G.
This means that the value

f 0 + g0
Z
1
2πi ∂G f +g
at t = 1 is the same as the value
f0
Z
1
,
2πi ∂G f
which yields Rouché’ theorem.
Rouché’s theorem. For meromorphic functions f and g on an open neigh-

borhood of a bounded domain G with piecewise smooth boundary such that
both f and g have no zeroes or poles on the boundary ∂G of G, if |f | > |g|
on ∂G, then the total number of zeroes of f + g in G minus the total number
of poles of f + g in G is equal to the total number of zeroes of f in G minus
the total number of poles of f in G .
Open Mapping Theorem for Holomorphic Function and Maximum Mod-

ulus Principle.
Open Mapping Theorem for Holomorphic Function. If f is a nonconstant

holomorphic function on connected open subset Ω of C, then the map Ω → C
defined by f is open (i.e., the image of an open set is open).
Proof. Since f 0 is not identically zero, the zeroes of f 0 are isolated. Let G
be an open subset of Ω and z0 be any point of G. If f 0 is nonzero at z0 , the
inverse mapping theorem guarantees that f can be locally inverted at z0 and
some open neighborhood Uz0 of z0 in G is mapped by f to some open subset
of C. If f 0 is zero at z0 , the vanishing order of f 0 at P is some positive integer
k. Write f (z) = f (z0 ) + (z − z0 )` g(z) with g(z) holomorphic and nonzero at
z0 . Then ` = k + 1.
Since g(z0 ) 6= 0, we can choose a holomorphic function h(z) on some

open neighborhood at z0 such that h(z)k+1 = g(z) (by using a branch of
1
log g(z) and setting h(z) = e k+1 log g(z) ). With respect to the local coordinate
ζ = (z − z0 )h(z) of Ω at z0 , the map f = f (z0 ) + ζ k+1 , as a power map
followed by a translation, maps some open neighborhood Uz0 of z0 in Gto an
open subset of C. f (G) = ∪P ∈G f (UP ) is open in C. Q.E.D.
Maximum Modulus Principle. If f is a nonconstant holomorphic function

on a connected open subset Ω of C such that the supremum M of |f | on Ω
is finite, then there does not exist any point P of Ω such that |f (P )| = M .
Proof. If such a point P exists, since f as a map is open, f (UP ) is an open

subset of C for some open neighborhood UP of P in Ω. Let D be the closed
disk in C with radius M and center 0. From |f (P )| = M = supΩ |f |, we get
the contradiction that f (UP ) is an open neighborhood of the boundary point
f (P ) of D and yet is contained in the subset f (Ω) of D.
Schwarz Lemma (Application of Maximum Modulus Principle).
Schwarz Lemma. Let f be a holomorphic function on the open unit disk D

such that |f | ≤ 1 on D and f (0) = 0. Then |f (z)| ≤ |z| on D and |f 0 (0)| ≤ 1.
If |f (z)| = |z| for some z ∈ D − {0} or |f 0 (0)| = 1, then f (z) = eiα z for some
α ∈ R.
The quotient function f (z) z

is holomorphic on D when its value at z = 0
is defined to be f 0 (0) at z = 0. On |z| ≤ r for any 0 < r < 1, it follows from
the maximum modulus principle that for |z| < r,

f (z)
≤ sup f (ζ) ≤ 1 .

z ζ r
|ζ|≤r
The conclusion follows from letting r → 1.
An important application of the Schwarz lemma is the following charac-

terization of biholomorphic maps of the open unit disk preserving the origin.
Any biholomorphic map f of D preserving the origin is of the form f (z) =

eiα z for some α ∈ R.
Let g be the inverse map of f . By the chain rule g 0 (0)f 0 (0) = 1. The
Schwarz lemma applied separately to f and g implies that |f 0 (0)| ≤ 1 and
|g 0 (0)| ≤ 1. Hence |f 0 (0)| = 1 and f (z) = eiα z for some α ∈ R by the Schwarz
lemma.
Application of Cauchy’s Residue Theory to Evaluation of Definite Inte-

grals and Representation of Functions as Infinite Sums of Partial Frac-
tions and as Infinite Products. Cauchy’s Residue theory can be applied to
evaluate certain definite integrals, especially the following types.
• Integral of a rational function of the sine and cosine functions over an
interval of length 2π.
• Integral of a rational function over the real line with the degree of the
denominator at least 2 more than that of the numerator.
• Integral of a product of either sine or cosine and a rational function with
real coefficients over the real line with the degree of the denominator
at least 1 more than that of the numerator.
• Integral involving branches of certain holomorphic functions.
• Expansion of meromorphic functions as infinite sums of partial frac-
tions.
• Expansion of entire functions as infinite products.
Integral of Rational Function of Sine and Cosine Functions over Interval

of Length 2π. For the evaluation of definite integrals of the form
Z 2π
R (cos θ, sin θ) dθ,
θ=0
where R (x, y) is a rational function (with the value of R (cos θ, sin θ) assumed
finite at every point θ ∈ [0, 2π]), The substitution
iθ z + z1 z − z1 dz
z = e , cos θ = , sin θ = , dθ =
2 2i iz
is used to convert the integral into a contour integral
Z
f (z)dz
|z|=1
of a meromorphic function f (z) of z over the contour |z| = 1, which can then
be evaluated by using Cauchy’s residue theory.
Integral of Rational Function over Real Line with Degree of Denom-

inator At Least 2 More Than That of Numerator. For the evaluation
of Z ∞
P (x)dx
,
x=−∞ Q(x)
where P (x) and Q(x) are polynomials with deg Q(x) ≥ deg P (x) + 2 and no
zero of Q(x) for real x, the meromorphic function used is
P (z)
Q(z)
over the contour of the boundary of the upper half disk of radius R centered
at the origin as R → ∞.
The condition of deg Q ≥ 2 + deg Q is needed to conclude that
Z ∞
P (z)dz
lim =0
R→∞ C
R
Q(z)

iθ
with CR = {Re 0 ≤ θ ≤ π}.
The formula is
Z ∞
P (x)dx X P (z)
= 2πi Resa .
x=−∞ Q(x) Im a>0
Q(z)
Integral of Product of Either Sine or Cosine Function and Rational

Function with Real Coefficients over Real Line with Degree of Denom-
inator At Least 1 More Than That of Numerator. For the evaluation
of Z ∞ Z ∞
P (x) cos xdx P (x) sin xdx
or ,
x=−∞ Q(x) x=−∞ Q(x)
where P (x) and Q(x) are polynomials of real coefficients with deg Q(x) ≥
deg P (x) + 1 and no zero of Q(x) for real x except simple ones at the zeros
of respectively cos x and sin x, the meromorphic function used is
P (z)eiz
Q(z)
over the contour of the boundary of the upper half disk of radius R centered
at the origin as R → ∞.
When deg P = 1 + deg Q, to conclude
P (z)eiz dz
Z
lim = 0,
R→∞ C
R
Q(z)
one needs to use integration by parts

z=R
P (z)eiz dz P (z)eiz d P (z) eiz
Z Z
= − dz
CR Q(z) iQ(z) z=−R CR dz Q(z) i
and the fact that on the right-hand side of
d P (z) P 0 (z)Q(z) − P (z)Q0 (z)

=
dz Q(z) Q(z)2
the degree of the denominator is at least 2 more than that of the numerator.
When there is a simple zero of Q(x) which is cancelled by the zeroes of

sine or cosine in the numerator, a detour of a small half-circle around the
simple zero has to be use to get the contribution of one half of the residue
from the simple zero of Q(x). At the end the real part and the imaginary
part of the result from residue theory. The final formula is the following.
Z ∞
P (x) cos x dx
x=−∞ Q(x)
!
X P (z)eiz X P (z)eiz
= Re 2πi Resa + πi Resa ,
Im a>0
Q(z) Im a=0
Q(z)
Z ∞
P (x) sin x dx
x=−∞ Q(x)
!
X P (z)eiz X P (z)eiz
= Im 2πi Resa + πi Resa .
Im a>0
Q(z) Im a=0
Q(z)
Definite Integrals Evaluated by Using Branches of Holomorphic Func-

tions. Some definite integrals which can be evaluated by the theory of
residues involve branches of holomorphic functions. For this type of definite

integrals there are no general formulas for evaluation. However, there are
some natural ways of choosing contours and branches for the meromorphic
functions used. Some examples of such definite integrals are
∞
(log x)2 dx
Z Z ∞
log x dx
, ,
x2 + 1 x4 + 1
Zx=0
∞
x=0
xα dx
with − 1 < α < 0,
1+x
Zx=0
1
dx
with 0 < α < 1.
x=0 xα (1 − x)1−α
For the evaluation of Z ∞

log x dx
x=0 x2 + 1
the contour used is the boundary of the indented upper half disk
n o
z ∈ C ε < |z| < R, Im z > 0

with radius R and an indentation of radius ε > 0. Then pass to limit as

R → ∞ and ε → 0. For the evaluation of
Z ∞ α
x dx
with − 1 < α < 0
x=0 1 + x
the contour used is the boundary of the domain which is the annulus
n o
z ∈ C ε < |z| < R

with inner radius ε and outer radius R minus the right half-strip
n o
z ∈ C Re z ≥ 0, −ε ≤ Im z ≤ ε

with width 2ε and axis Im z = 0. Then pass to limit as R → ∞ and ε → 0.
The meromorphic function used involves choosing branches for the loga-
rithmic function and the fraction power function which can also be expressed
in terms of the logarithmic function.
Branches of Multivalued Holomorphic Functions. Some holomorphic

functions such as the logarithmic function and the fractional power func-
tion are multi-valued. A continuous function obtained by choosing one value
of a multi-valued function f for each point of a domain D is called a branch of
the multi-valued function of f on D. The domain D usually is constructed by
removing certain line-segments called “branch-cuts” from the natural max-
imum domain of definition for the multi-valued function. The value of the
branch function at a point approaching one side of a branch-cut differs from
the value when the point approaches from the other side. That is the rea-
son why the branch-cut is needed to guarantee that the branch function is
a continuous function. For example, for log z a branch can be defined on
C − [0, ∞) by choosing arg z to be in (2kπ, (2k + 1)π) for any k ∈ Z.
For the product of the branches of two different multi-valued functions,

a line-segment common to both branch-cuts may become unnecessary for
the product if the two discrepancies of values across it for the two branch
functions cancel out. An example is the function z α (1 − z)1−α (with 0 < α <
1) for which the branch-cut of [0, ∞) is used to construct a branch-function
for the factor z α while the branch-cut of [1, ∞) is used to construct a branch-
function for the factor (1 − z)1−α . Across the common part (1, ∞) of two
branch-cuts, the two discrepancies of values for the two branch functions
cancel out so that a branch of the multi-valued function z α (1 − z)1−α can be
defined on C − [0, 1].
Partial Fraction Expansion of Meromorphic Functions (Application of

Cauchy’s Integral Formula to Meromorphic Function after Modifica-
tion to Ensure Vanishing of Limit Contour Integral). For a holomorphic
function f on an open neighborhood of a bounded domain G with piecewise
smooth boundary, Cauchy’s integral formula
Z
1 f (ζ)dζ
f (z) =
2πi ∂G ζ − z
reproduces f (z) from an integral. Suppose f (z) is now assumed to be mero-

morphic instead of holomorphic. Cauchy’s integral formula will not hold in
general. There will be some additional terms from the contribution of the
pole singularities of f . The additional terms are some partial fractions. To
represent f as an infinite sum of partial fractions, one would like to get rid
of the integral over ∂G by letting ∂G recede to infinity and at the same time
keep these additional partial fraction terms from the pole singularities of f .
The integral over ∂G would go to 0 as ∂G recedes to infinity if certain

growth order conditions are met. To get to such a situation, we replace the
integrand
f (ζ)dζ
ζ −z
by
f (ζ)dζ
ζ p+1 (ζ − z)
and impose the growth condition f (z) = o (Rnp+1 ) on a contour Cn , where Rn
is the distance from the origin, when ∂G is replaced by Cn whose length is of
order O(Rn ) with Rn → ∞ as n → ∞. We formulate in the following state-
ment the result obtained for the partial fraction expansion of meromorphic
functions.
Theorem. Suppose f (z) is a meromorphic function on C whose poles are

simple {an }1≤n<∞ with 0 < |a1 | ≤ |a2 | ≤ · · · so that the residue of f (z) at
an is bn . Suppose that there is a sequence of closed contours Cn such that
Cn includes a1 , · · ·, an but no other poles. Assume that the distance Rn from
Cn to the origin goes to infinity as n → ∞ and the length Ln of Cn is of the
order O(Rn ). Assume that on Cn we have f (z) = o(Rnp+1 ). Then
p ∞
zν z2 zp

X
(ν)
X 1 1
f (z) = f (0) + bn + + 2 + · · · + p+1 .
ν=0
ν! n=1
z − an an an an
To prove it, we apply the theorem of residue to the integral

Z
1 f (ζ)
p+1
dζ.
2πi Cn ζ (ζ − z)
The residue at ζ = 0 is obtained by expanding fζ−z

(ζ)
in Laurent series in ζ
around ζ = 0. We have
ζ2

f (ζ) −1 1 ζ 0 1 00 2
= p+1 + + + ··· f (0) + f (0)ζ + f (0)ζ + · · ·
ζ p+1 (ζ − z) ζ z z2 z3 2
1
and the coefficient of ζ
is
f (0) f 0 (0) f (p) (0)

1
− + + · · · + .
z zp z p−1 p!
The residue at ζ = z is given by zfp+1 (z) bn
. The residue at an is ap+1 (a . Since
n n −z)
as n → ∞ the integral becomes zero, we get
∞
1 f (0) f 0 (0) f (p) (0)

f (z) X bn
− + + · · · + + + p+1 = 0,
z zp z p−1 p! z p+1 n=1 an (an − z)
which means that
∞
z p (p) X bn z p+1
f (z) = f (0) + z f 0 (0) + · · · +
f (0) +
p! ap+1 (an − z)
n=1 n
p ∞
z ν (ν) z2 zp

X X 1 1
= f (0) + bn + + + · · · + p+1 .
ν=0
ν! n=1
z − an an a2n an
The last expression comes from writing
z p+1 = (z p+1 − ap+1 p+1
n ) + an
and then factoring

p
X
z p+1 − ap+1
n = (z − an ) z ν ap−ν
n .
ν=0
Note that
1 z2 zp
+ 2 + · · · + p+1
an an an
is the p-th Taylor polynomial for the function
1
z − an
at z = 0 from geometric series expansion.
An Example of Partial Fraction Expansion. The meromorphic function is

f (z) = cot z − z1 and Cn is the square with corners at (n + 21 )(±1 ± i). When
|y| > π2 we have
2iz
e + 1 e2y + 1

2 2
| cot z| ≤ 2iz
≤
2y
= 1 + 2y ≤1+ π
e −1 e −1 e −1 e −1
and hence uniformly bounded. The function cot z is bounded on the line
joining 21 (1 − i)π to 12 (1 + i)π and we can use periodicity
P and conclude that
cosec z is uniformly bounded on Cn . Hence cot z − z1 = n∈Z−{0} z−nπ1 1
+ nπ
because the residue of cot z at nπ is 1.
Infinite Product Expansion. From an infinite product expansion, one can

get a partial fraction expansion by taking the derivative of the logarithm
of the infinite product. This process can be reversed to yield an infinite
product expansion of an entire function from a partial fraction expansion of
the logarithmic derivative of the entire function. As an example, we derive the
infinite product expansion of sin πz by using the partial fraction expansion
d
of π cot πz = dz log sin πz. By replacing z by πz in the above partial fraction
expansion, we get
1 X 1 1

π cot πz = + + ,
z z−n n
n∈Z−{0}
from which the infinite product expansion of sin πz will derived.
We are going to exponentiate the indefinite integral of

X 1
d sin πz 1
log = π cot πz = + .
dz πz z−n n
n∈Z−{0}
Since
d z z 1 1
log 1 − en = + ,
dz n z−n n
when we integrate from z = 0 to z, we get
Z ζ=z
1 1 z z
+ dζ = log 1 − en .
ζ=0 ζ −n n n
After exponentiating, we obtain
sin πz Y z z
=C 1− en .
πz n
n∈Z−{0}
Passing to limit as z → 0, we determine C to be 1 and get

Y z z
sin πz = πz 1− en .
n
n∈Z−{0}
Evaluation of Sum of Values of Rational Functions at Integral Points

by Residues and Cotangent Function. Suppose f (z) is a rational function
whose poles are simple nonintegers a1 , · · · , ak with residues b1 , · · · , bk such
that the degree of the denominator of f is at least two more than that of
its numerator. Let Cn be the square with corners at (n + 21 )(±1 ± i). The
integral Z
π cot πz f (z)dz
Cn
goes to zero as n → ∞. The use residues in evaluating
Z
π cot πz f (z)dz
Cn
and passing to limit as n → ∞ yields
Theorem. For a rational function with simple poles at nonintegers a1 , · · · , ak

with residues b1 , · · · , bk ,
∞
X k
X
f (n) = −π bν cot π aν .
n=−∞ ν=1
If we use cosec πz instead of cot πz we can obtain the sum of the series
P ∞ n
n=−∞ (−1) f (n).
Theorem. For a rational function with simple poles at nonintegers a1 , · · · , ak

with residues b1 , · · · , bk ,
∞
X k
X
n
(−1) f (n) = −π bν cosec π aν .
n=−∞ ν=1
As an example, we sum the series

∞
X 1
n=−∞
n2 + a2
for a > 0 by using the rational function

1
f (z) =
z2 + a2
which has simple poles at z = ai and at z = −ai with residues respectively

−i
2a
and 2ai to get
∞
X 1 −i cot πai i cot π (−ai) π
= −π + = coth πa.
n=−∞
n + a2
2 2a 2a a
This concludes the review of basic complex analysis

We will start the discussion of elliptic and theta functions in the next lecture

Review Basic Complex Analysis Complete

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Review Basic Complex Analysis Complete

Uploaded by

Copyright:

Available Formats

Math Math 213a (Fall 2021) Yum-Tong Siu 1

Review of Basic Complex Analysis

Complex Derivative and Cauchy-Riemann Equations. For a complex-

Cauchy-Riemann Equations not Sufficient Condition for Existence of

Cauchy-Riemann equations for f (z) = u(x, y) + iv(x, y) at a point c = a + bi

Necessary and Sufficient Condition for Existence of Complex Derivative

v(x, y) = v(a, b) + vx (a, b)(x − a) + vy (a, b)(y − a) + E2 (x, y)

Proof. The existence of the complex derivative of f (z) = u(x, y) + iv(x, y)

Taking the real and imaginary parts of

v(x, y) = v(a, b) + vx (a, b)(x − a) + vy (a, b)(y − a) + Im E(z).

On the other hand, when

u(x, y) + iv(x, y) = (u(a, b) + iv(a, b))

Geometric Interpretation of Cauchy-Riemann Equations. The condition

v(x, y) = v(a, b) + vx (a, b)(x − a) + vy (a, b)(y − a) + E2 (x, y)

which is the tangent map (or derivative) of Φ at (x, y) = (a, b).

When the R-vector space R2 is identified with C-vector space C, the

means that the map f can be approximated at z = c, to an order higher than

(where ∂R means the boundary of R in the counterclockwise sense) which is

Proof. The theorem of Cauchy simply follows from Stokes’s theorem

The key additional assumption here is the existence and continuity of ux ,

Theorem of Cauchy-Goursat for a Rectangle. If the complex derivative

Before giving a rigorous argument, let us first intuitively understand the

For a small closed rectangle Rc containing c of total boundary length `c ,

by Cauchy’s theorem applied to Pc (z), with

From the above argument applied to each Rcj to get

Continue this process to end up with a sequence of closed subrectangles R(m)

Now use compactness of R to conclude that there is some c ∈ R which is in

when ε`2 < A and m ≥ mε .

Theorem of Cauchy-Goursat (for Domain with Piecewise Smooth Bound-

where Cz is any broken path joining z0 to z with each line-segment either

The theorem of Cauchy-Goursat for a closed rectangle guarantees that

Theorem of Cauchy-Goursat. If f (z) is a holomorphic function on an open

as a Riemann integral requires continuity of ϕ0 (t). The boundary of G is

comes from the application of the fundamental theorem of calculus which

In the case of more than one component in the boundary of G, curve-

This explicit computation is used in the proof of Cauchy’s integral formula

Cauchy’s Integral Formula. If f (z) is a holomorphic function on an open

for 0 < ρ ≤ ρ0 , which means that the integral

Differentiation of Definite Integral with respect to a Real Parameter of

For c0 ∈ (c, d), by Fubini’s theorem of exchanging the order of integration

Cauchy’s Integral Formula for Derivatives of Holomorphic Functions.

is infinitely differentiable as a function of x and y (where z = x + iy) and is

The infinite differentiability of h follows from

for p, q ∈ N ∪ {0}, which is obtained by applying the above result on differ-

We are now ready to prove Cauchy’s integral formula for derivatives of

by induction on the nonnegative integer n. The case of n is simply the

a function of x, y and is holomorphic on G. Differentiating both sides of (\)n

Uniform Limit of Holomorphic Functions. From the Cauchy integral for-

Cauchy’s Kernel. Cauchy’s integral formula

Relation Between Cauchy’s Kernel and Newton’s Potential. The Cauchy

Consider the complex-valued differential dz = dx + idy and dz̄ = dx − idy

The Cauchy-Riemann equations ∂f ∂x

for the Newton potential should be written as

because the right-hand side yields a number when evaluated at a function.

Cauchy Integral Formula for Smooth Functions. After a translation and

Suppose G is a bounded domain in C with piecewise smooth boundary and

Cauchy’s Integral Formula for Smooth Functions as Complex-Analytic

which we can rewrite as

The way to do it without using some other solution F (z) is to apply

Power Series and Laurent Series Expansion of Cauchy Kernel. Since

converges and gives a power series expansion of the Cauchy kernel.

converges and gives a Laurent series expansion of the Cauchy kernel

Power Series Expansion of Holomorphic Function on Disk. Suppose f is