Complex Variables Formula Sheet

William Bevington

Chapter One - Holomorphicity Theorem. If S ⊆ C is compact (i.e. closed and bounded) then f (S) is
compact.
Theorem (Triangle Inequality).
Definition. A function f : C → C is differentiable at z0 if
|z + w| ≤ |z| + |w| limz→z0 f (z)− f (z0 )
z−z0 = 0. Furthermore differentiability gives continu-
ity.
||z| − |w|| ≤ |z − w|
Theorem. Let z0 = x0 + iy0 ∈ C and U ⊆ C be a neighbourhood of z0
Definition. The argument of z ∈ C is arg(z) := {θ : z = |z|eiθ }, the with f : U → C, f = u + iv differentiable at z0 . The Cauchy-Riemann
principle argument is Arg(z) ∈ arg(z) ∩ (−π, π] and is unique. equations are
∂u ∂v ∂v ∂u
Theorem (1.1.19). Let z, w ∈ C be non-zero. Then arg(zw) = arg(z) + = and =− .
∂x ∂y ∂x ∂y
arg(w) and Arg(zw) = Arg(z) + Arg(w).
Definition. A function f : C → C is holomorphic at z0 if it’s differ-
Definition. The open (resp. closed) ε-disk centred at z0 is Dε (z0 ) := entiable on an open neighbourhood of z .
0
{z ∈ C : |z − z0 | < ε} (resp. Dε (z0 ) := {z ∈ C : |z − z0 | ≤ ε}). The
puntured disk is D0ε (z0 ) := Dε (z0 ) \ {z0 }. Definition. If u : U ⊆ R2 → R is harmonic (i.e. uxx + uyy = 0) and
f = u + iv is holomorphic, then v is the harmonic conjugate of u.
Definition. A subset D ⊆ C is open if ∀z ∈ D : ∃ε > 0 : Dε (z) ⊆ D (or
it’s a union of open disks) and is closed if C \ D is open. If z ∈ D is Definition. The complex exponential is exp(z) = ex (cos(y)+i sin(y))
open we say D is a neighbourhood of z. where z = x + iy. It is holomorphic on all of C (prop 1.6.2).

Definition. Let S ⊆ C then z0 ∈ C is a limit-point of S if ∀ε > 0 : Theorem. Let z, w ∈ C, then

D0ε (z0 ) ∩ S 6= 0.
/ If LS is the set of limit points of S then S := S ∪ LS is
the closure of S. exp(z + w) = exp(z) exp(w) and exp(z + 2πi) = exp(z).

Theorem (1.2.9). A complex sequence zn converges iff Re(zn ) and Definition. The complex logarithm for z ∈ C is log(z) := {w ∈ C :
Im(zn ) converge. exp(w) = z}.

Theorem. The complex plane C is complete, namely zn is convergent Theorem (1.7.3). Let z, w ∈ C, then
⇔ zn is Cauchy.
log(z) = ln |z|+i arg(z), log(zw) = log(z) + log(w),
Theorem (Bolzano-Weierstrass). If zn is a bounded sequence then it and log(1/z) = − log(z).
has a convergent subsequence.
Definition. The principle logarithm is Log(z) := ln |z| + iArg(z).
Theorem (1.3.3). Let f : S ⊆ C → C and z0 = x0 + iy0 ∈ S and z =
x + iy, a0 ∈ C, then ∃u(x, y), v(x, y) such that f (z) = u(x, y) + iv(x, y). Definition. A branch cut is Lz0 ,θ := {z ∈ C : z = z0 + reiθ , r ≥ 0},
Then a0 = limz→z0 f (z) iff Re(z) = lim(x,y)→(x0 ,y0 ) u(x, y) and Im(z) = giving the cut plane D0,π := C \ L0,π . If we let Argθ (z) := arg(z) ∩
lim(x,y)→(x0 ,y0 ) v(x, y). (θ , θ + 2π] then Logθ := ln |z| + iArgθ (z).

Definition. A function f : C → C is continuous if f −1 (U) is open for Theorem (1.7.10). Let θ ∈ R and U ⊆ C with g : U → C holomor-
all U ⊆ C, equally if ∀ε > 0 : ∃δ > 0 : | f (z) − f (z0 )| < ε whenever phic, then Log(g(z)) is holomorhpic on U ∩ g−1 (D0,θ ). Particularly, if
|z − z0 | < δ then f is continuous at z0 . g is hol. on C then Log(g(z)) is holomorhpic on g−1 (D0,θ ).

.................................................................................................................................

Chapter Two - Möbius transformations verse bijections (ϕ : S2 → C̃, ψ : C̃ → S2 ) given by:

|z|2 − 1
 
Theorem. Let f : U → C be holomorphic with U ⊆ C open, then f is X + iY 2x 2y
ϕ(X,Y, Z) := and ψ(z) := , ,
conformal (preserves angles) at z0 ∈ U iff f 0 (z0 ) 6= 0. 1−Z |z|2 + 1 |z|2 + 1 |z|2 + 1

Definition. A Möbius transformation is any function of the form Theorem (2.4.3). A Möbius transformation maps circlines (circles
∼
cz+d where ad − bd 6= 0. This forms a group M = SL(2; C).
f (z) = ax+b and lines) to circlines.

Definition. The Riemann sphere is S2 := {(X,Y, Z) ∈ R3 : X 2 +Y 2 + Theorem. The cross-ratio is the unique Möbius transformation which
Z 2 = 1}. The extended complex plane is C̃ := C ∪ {∞} where ∞1 = 0 sends (z2 , z3 , z4 ) 7→ (1, 0, ∞). The image of z under this is given by
and z · ∞ = ∞ · z = ∞ = 01 for any z ∈ C.
z − z3 z2 − z4
[z, z2 , z3 , z4 ] = .
Theorem. Möbius transformations are holomorphic and conformal z − z4 z2 − z3
on C̃.
Theorem (2.5.7). Let M be a Möbius transformation, then
Theorem. Let z = x + iy. Stereographic projection is the pair of in- [Mz1 , Mz2 , Mz3 , Mz4 ] = [z1 , z2 , z3 , z4 ].

1
Chapter Three - Complex integration Definition. Let Γ be a loop, then Int(Γ) is the interior, and Ext(Γ) is
the exterior so that C = Int(Γ) ∪ Γ ∪ Ext(Γ).
Definition. Let [a, b] ⊆ R be a closed interval and f : [a, b] → C be
of the form Rf = u + iv, then f is integrable if u and v are in the real Definition. A domain D is simply-connected if for any loop Γ :
R R
sense. Then [a,b] f (t)dt = [a,b] u(t)dt + i [a,b] v(t)dt. Int(Γ) ⊆ D.
R b dF Theorem. Cauchy-Integral: Let Γ be a loop and f be holomorphic
Theorem (3.1.2). Integration in C is linear, and a dt dt = F(b) −
Rb Rb inside and on Γ, then the following hold:
F(a). One estimate for integration is a f (t)dt ≤ a | f (t)|dt.
Z
Definition. A parametrized curve Γ from z0 to z1 (distinct) is a con- f (z)dz = 0,
Γ
tinuous function γ : [t0 ,t1 ] → C with γ(t0 ) = z0 and γ(t1 ) = z1 . It is
1 f (z)
Z
regular if γ 0 (t) exists, is continuous and non-zero. dz = f (z0 ),
2πi Γ z − z0
Definition. Let Γ be a regular curveRand f : Γ → C continuous. The n!
Z
f (w)
integral of f along Γ is Γ f (z)dz = tt01 f (γ(t))γ 0 (t)dt.
R dz = f (n) (z).
2πi Γ (w − z)n+1
R t1
Definition. The arc-length of Γ is l(Γ) := t0 |γ 0 (t)|dt.
Theorem (3.4.11). Let Γ be a loop not passing through z0 , then
Theorem (3.2.9,ML Lemma). Let Γ be regular and f : Γ → C con- (
tinuous, then Z
1 2πi if z0 ∈ Int(Γ)
Z dz = .
f (z)dz ≤ max | f (z)|l(Γ) Γ z − z0 0 otherwise
Γ z∈Γ

Definition. D ⊆ C is a domain if it’s open and ∀z, w ∈ D : ∃Γ, a con- Theorem

R
(3.4.12).R Let Γ1 , Γ2 be loops with f holomorphic on both
tour connecting z to w. then Γ1 f (z)dz = Γ2 f (z)dz, i.e. the two loops can freely be deformed
into each other.
Theorem. Fundamental Theorem of Calculus: Let D be a do-
main
R
and Γ ⊆ D a contour connecting z0 , z1 ∈ D and F 0 = f then Theorem (3.5.2). Let f be holomorphic on a domain D, then f has
Γ f (z)dz = F(z1 ) − F(z0 ). infinitely many derivatives, all of which are holomorphic.
Definition. Let Γ ⊆ D be a contour in domain D ⊆ C, Γ is a closed
contour if it has equal endpoints (γ(t0 ) = γ(t1 )). Theorem (Morera). Let D ⊆ C be a domain and f is continuous with
R
Γ f (z)dz = 0 for all loops Γ, then f is holomorphic on D.
Theorem. Path-independence: Let D ⊆ C be a domain with contin-
uous f : D → C then the following are equivalent: Theorem. Let f : DR (z0 ) → C be holomorphic and bounded by M.
Then
• f has an anti-derivative F on D, n!M
| f (n) (z0 )| ≤ n .
•
R
f (z)dz = 0 for all closed contours Γ ⊆ D, R
Γ
R
• all contour integrals Γ f (z)dz are independent of path, thus de- Theorem. Liouville: Let f be holomorphic on C and bounded, then
pend only on the end-points. f is constant.

Definition. A contour Γ is simple if it has no self-intersections, if it Theorem. Maximum modulus principle: Let D ⊆ C be a domain on
is also closed then we call it a loop. A loop is positively-oriented if a which f is holomorphic and bounded by M. If f achieves its maximum
parametrisation γ goes around anti-clockwise. inside D the f is constant on D.

2
Chapter Four - Series expansions Definition. The Laurent series expansion of a function f is
∑∞j=−∞ a j (z − z0 ) j = ∑∞j=0 a j (z − z0 ) j + ∑∞j=1 a− j (z − z0 )− j .
Theorem. Convergence tests
Definition. The open annulus of radii r and R is Ar,R (z0 ) = DR (z0 ) \
• Comparison test: Suppose ∀n : |zn | ≤ Mn with ∑∞j=0 M j conver- Dr (z0 ).
gent, then ∑∞j=0 z j converges.
Theorem. The coefficients of a Laurent series for a holomorphic func-
• The series ∑∞j=0 c j converges iff |c| < 1. tion f are given by
• Ratio Test: Let L := limn→∞ | f raczn+1 zn |, then zn converges if 1
Z
f (z)
L < 1 and diverges if L > 1. aj = j+1
dz,
2πi Γ (z − z0 )

• Weierstrass M: If fn is a sequence of functions ∀ j : | f j | ≤ M j

for Γ ∈ Ar,R (z0 ).
and ∑∞j=0 M j converges then fn converges uniformly.
Definition. Let ( fn )n∈N be a sequence of functions, fn converges Theorem (4.4.7). The Laurent series expansion of holomorphic f is
pointwise to f if ∀ε > 0 : ∃N ∈ N : ∀n ≥ N : | fn (z) − f (z)| < ε. unique.

Definition. Let ( fn )n∈N be a sequence of functions, fn converges uni- Definition. We say z0 is a singularity of f if f isn’t holomorphic at
formly to f if ∀ε > 0 : ∃N ∈ N : ∀n ≥ N : ∀z : | fn (z) − f (z)| < ε. z0 . It is isolated if ∃R > 0 : f is holomorphic on D0 (z0 ), and of order
m if f (z0 ) = · · · = f (m−1) (z0 ) = 0 6= f (m) (z0 ).
Theorem (4.1.21,4.1.22). If fn converges uniformly we may commute
limits with integrals (4.1.21) and integrals with sums (4.1.22). Theorem (4.5.5). If zn → z0 and ∀n : zn ∈ D which is a neighbourhood
domain of z0 and f (z0 ) = 0 then f (z) = 0 for all z ∈ D.
Theorem (4.1.23). If every fn is holomorphic and fn → f uniformly
then f is holomorphic. Definition. Let z0 be a singularity of a function f , then
Theorem (4.2.2). Let P = ∑∞j=0 a j (z − z0 ) j be a power-series then • z0 is removable if ∀ j < 0 : a j = 0,
∃R ∈ [0, ∞] : (called the radius of convregence) such that:
• of order m if ∀ j < −m : a j = 0 but a j 6= 0,
• P converges on DR (z0 ),
• essential if there are infinitely many a j 6= 0 with j < 0.
• P converges uniformly on Dr (z0 ) for any r < R,
• P diverges on C \ DR (z0 ). Theorem (4.5.8). Let f = ∑∞j=0 a j (z−z0 ) j have removable singularity
z0 then re-defining f (z0 ) = a0 makes f holomorphic at z0 .
aj
Theorem (4.2.4). If r := lim j→∞ a j+1 exists then it’s the radius of
Theorem (4.5.11). Let f , g be holomorphic at z0 with z0 a zero of
convergence of ∑∞j=0 a j (z − z0 ) j . order m of g then
Theorem (4.2.6). The power series ∑∞j=0 a j (z − z0 ) j with radius of • if z0 isn’t a zero of f then f /g has a pole of order m at z0 ,
convergence R is holomorphic on DR (z0 ).
• if z0 is a zero of order k of f then f /g has a pole of order m − k
Definition. The Taylor seires of f , for holomorphic f , is
at z0 if m > k and removable singularity otherwise.
∞
f ( j) (z0 )
∑ (z − z0 ) j . Definition. We say F : D̃ → C is an analytic continuation of f : D →
j=0 j! C with D̃ ⊆ D ⊆ C if F(z) = f (z) for z ∈ D and F is holomorphic.
Theorem (4.3.2). If f is holomorphic on DR (z0 ) then it admits a Tay- Theorem (4.6.4). Identity theorem: Let D ⊆ C be a domain with f
lor ∑∞j=0 a j (z − z0 ) j series which converges uniformly with radius of holomorphic on D and f (z) = 0 for all z ∈ DR (z0 ) ⊆ D then f (z) = 0
convergence R. for all z ∈ D.
Definition. A function f is analytic if it admits a convergent power- Theorem (4.6.5). Let D ⊆ C be a domain with f , g : D → C holomor-
series. phic with ∀z ∈ DR (z0 ) : f (z) = g(z) then f (z) = g(z) for all z ∈ D.
Theorem (4.3.5). Every holomorphic function is analytic.
Theorem (4.6.7). Let zn → z0 and ∀n : f (zn ) = 0 with f : D → C holo-
Theorem (4.3.9). The Taylor series of f 0 (z)
is the term-by-term morphic, then ∀z ∈ D : f (z) = 0.
derivative of the Taylor series of f (z) (since Taylor series converge
uniformly). Theorem (4.5.8). Let D ⊆ C be a domain with f , g : D → C holomor-
phic with and zn → z0 , f (zn ) = g(zn ) then ∀z ∈ D : f (z) = g(z) (use
Theorem (4.3.12). Taylor series are unique, specifically; the Taylor this to prove that sin2 (z) + cos2 (z) = 1 holds for complex sin, cos since
series of a function is equal to any valid power-series. f = g on the real axis).

NOTE:
∞ ∞
z2 j+1
∑ z j is convergent on D1 (0). sin(z) = ∑ (2 j + 1)!
j=0 j=0
∞
z2 j ∞
zj
cos(z) = ∑ (2 j)! exp(z) = ∑
j=0 j=0 j!

3
Chapter Five - Residue calculus where N0 ( f ) = ∑lj=1 ord(w j ) is the sum of the orders of the zeros of
f and N∞ ( f ) = ∑kj=1 ord(z j ) is the sum of the orders of the poles of f
Definition. The residue of a function f at isolated singularity z0 is (the number of poles in Int(Γ), counted with multiplicity).
1
Res( f , z0 ) = a−1 , the coefficient of z−z 0
in the Laurent series expan-
sion of f . Theorem. Rouché’s Theorem: Let Γ be a loop and f , g be holomor-
phic inside and on Γ with
Theorem (5.1.4). Let f be holomorphic on D0R (z0 ) with removable
singularity z0 , then Res( f , z0 ) = 0. ∀z ∈ Γ : | f (z) − g(z)| < | f (z)|
Theorem (5.1.5). Let f be holomorphic on D0R (z0 ) where z0 is a pole then N0 ( f ) = N0 (g).
of order m, then
Theorem. Open-Mapping theorem: Let D ⊆ C be a domain and f is
1 d m−1
Res( f , z0 ) = lim ((z − z0 )m f (z)). non-constant and holomorphic on D, then the image f (D) is open.
z→z0 (m − 1)! dzm−1
Theorem. Maximum Modulus: Let D ⊆ C be a domain and f be
Theorem (5.1.7). Let g, h be holomorphic on D0R (z0 ) where z0 is a holomorphic and non-constant, then | f (z)| doesn’t attain its maximum
simple zero of h and g(z0 ) 6= 0, then on D.
g(z0 ) Theorem (5.2.18). Suppose f is holomorphic on domain D, if any
Res( f , z0 ) = .
h0 (z0 ) of Re( f ), Im( f ), | f |, or Arg( f ) are constant functions then f is also
constant.
Theorem. Cauchy Residue Theorem: Let Γ be a loop with f holo-
morphic on Int(Γ) \ {z1 , . . . , zk } for isolated singularities z1 , . . . , zk . Theorem. Jordan Lemma: Let P/Q be rational with deg(Q) ≥
Then deg(P) + 1 then
Z k
f (z)dz = 2πi ∑ Res( f , z j ). (
γ j=1 P(z) Cρ+ for a > 0
Z
lim exp(iaz) dz = 0 where C =
Definition. A function f is meromorphic on a domain D if ∀z ∈ D, f ρ→∞ C Q(z) Cρ− for a < 0
has a pole of finite order or is holomorphic.
Theorem (5.5.3). Let D be a domain with f meromorphic on D with
Theorem. The Argument Principle: Let Γ be a loop in C and f mer- simple pole c ∈ D, if γ : [θ0 , θ1 ] ⊆ [0, 2π] → C, θ 7→ c + r exp(iθ )
emorphic on Int(Γ) and holomorphic on Γ, then parametrizes the arc Sr then
1 f 0 (z)
Z Z
dz = N0 ( f ) − N∞ ( f ), lim f (z)dz = i(θ1 − θ0 )Res( f , c).
2πi Γ f (z) r→0+ Sr

.................................................................................................................................

Sample questions W ORKSHOP 4; Q UESTION 6: The function f (z) = (z2 + 1)1/2 has
branches given by exp( 21 g(z)) for g(z) ∈ log(z2 + 1) = log(z + i) +
For infinite series: log(z − i) a branch og the logarithm. Thus the branch of f holomorh-
∞ pic on the unit disc D1 (0) is given by:
cot(πz) 1
Z
π
dz = ∑ Res( f , n) = ∑ r2 = 6 .
Γ z2 n j=0
 
1 
f (z) = exp Log−π/2 (z + i) + Logπ/2 (z − i) .
2
For finding series expansions:
∞ W ORKSHOP 5; Q UESTION 2:
1
1 − f (z)
= ∑ f (z) j for | f (z)| < 1,
j=0
z−i
1 d 1 f1 (z) = : U1 = {z ∈ C : Im(z) > 0} → D1 (0),
= z+i
(1 − z)2 dz 1 − z exp(πz) − i
∞  j f2 (z) = : U2 = {z ∈ C : 0 < Im(z) < 1} → D1 (0),
1 1 1 k exp(πz) + i
= · = ∑ z .
z−k z 1 − k/z j=0 z2 − i
f3 (z) = : U3 = {z ∈ C : Im(z), Re(z) > 0} → D1 (0),
z2 + i
For maps to the unit disc: The function
z2 − 1 n  π π o
z−1 f4 (z) = 2 : U4 = z ∈ C : Arg(z) ∈ − , → D1 (0).
f (z) = , z +1 4 4
z+1
maps the imaginary axis (equation |z − 1| = |z + 1|) to the unit circle.
On any disc: On z ∈ Dr (z0 ) notice that zz = |z|2 =⇒ z = rz and
2 W ORKSHOP 6; Q UESTION 3:
so Z Z
z + z z + r2 /z R3 f (z)dz 6= | f (z)|dz ∈ C
Re(z) = = .
2 2 Γ Γ

4
Useful Formulae

sin(iz) = i sinh(z)
sin(a ± b) = sin(a) cos(b) ± cos(a) sin(b) cos(iz) = cosh(z)
cos(a ± b) = cos(a) cos(b) ∓ sin(a) sin(b) cosh2 (z) − sinh2 (z) = 1
exp(iz) − exp(−iz)
sin(z) =
2i
exp(iz) + exp(−iz)
cos(z) =
2

Trigonometric Integral example

Theorem. Z 2π
R(cos θ , sin θ )
Z
R(cos θ , sin θ )dθ = f (z)dz, where f (z) = .
0 Γ i exp(iθ )
Example. Consider the integral
sin2 θ
Z 2π
I= dθ .
0 5 + 4 cos θ
z+z z+ 1z
Noting that z = eiθ = cos θ + i sin θ and so cos θ = Re(eiθ ) = 2 which on the unit circle becomes cos θ = 2 (since 1 = |z|2 = zz), then:

1 1 2



1 2i z − z i (z2 − 1)2
f (z) = 1 1

= ,
iz 5 + 4 · 2 z + z 8 z2 (z + 21 )(z + 2)

with poles at − 12 , −2 and 0, of which only 0 and − 12 lie inside D1 (0) and so we calculate their residues:
!
d i (z2 − 1)2 −5i
Res( f , 0) = lim =
z→0 dz 8 (z + 21 )(z + 2) 16
d i (z2 − 1)2
 
3i
Res( f , 1/2) = lim 2
=
z→ 12 dz 8 z (z + 2) 16

giving that  
−5i 3i π
I = 2πi ∑ z j = 2πi + = .
z j is a pole 16 16 4