Complex Functions and Analyticity

Complex functions map subsets of

C

C

; a function is analytic (holomorphic) at a point if it is complex-differentiable in a neighborhood of that point. The key result: complex differentiability is enormously stronger than its real counterpart — a once-differentiable complex function is automatically infinitely differentiable and equals its own Taylor series. Applies throughout physics (fluid flow, electrostatics), engineering (signal processing), and number theory. Foundations laid by Cauchy, Riemann, and Weierstrass in the 19th century.

A complex function $f : Ω \to C$ , where $Ω \subseteq C$ , assigns to each $z = x + i y$ a complex value $f (z) = u (x, y) + i v (x, y)$ . It is called analytic (synonymously holomorphic) on an open set if the limit defining its derivative exists at every point. This single requirement — that the derivative not depend on the direction of approach in the plane — forces the real and imaginary parts to satisfy the Cauchy–Riemann equations, links them to harmonic functions, and unleashes the entire edifice of complex analysis. The surprising fact: unlike real calculus, where a function can be differentiable once but not twice, complex differentiability at every point of an open set guarantees the function is infinitely differentiable and locally represented by a convergent power series. Analyticity is thus a rigidity property: the values of an analytic function on a tiny region determine it everywhere on a connected domain.

The Core Insight

In $R^{2}$ , the derivative of a map is a linear transformation that can stretch, shear, and rotate independently in different directions. Complex differentiability demands that this linear map be multiplication by a single complex number — equivalently, a rotation composed with a uniform scaling. This geometric constraint (angle-preservation, conformality) is the source of every deep theorem in the subject.

Formal Statement

Complex Function

A complex function is a map $f : Ω \to C$ where $Ω \subseteq C$ . Writing $z = x + i y$ with $x, y \in R$ , and decomposing the output as $f (z) = u (x, y) + i v (x, y)$ , we obtain two real-valued functions $u, v : Ω \to R$ called the real part $u = Re f$ and imaginary part $v = Im f$ .

Complex Derivative & Analyticity

Let $Ω$ be open and $z_{0} \in Ω$ . The function $f$ is complex-differentiable at $z_{0}$ if the limit

f^{'} (z_{0}) = h \to 0 lim \frac{f ( z _{0} + h ) - f ( z _{0} )}{h}

exists, where $h \in C ∖ {0}$ and $h \to 0$ along any path in the plane. The value $f^{'} (z_{0}) \in C$ is the derivative. The function $f$ is analytic (holomorphic) on $Ω$ if $f^{'} (z)$ exists at every point of $Ω$ . If $f$ is analytic on all of $C$ it is called entire.

⟳This mirrors the real difference quotient, but

h

ranges over a 2D neighborhood of

0

; the limit must be independent of the direction

ar g h

— the crux of the entire theory.

Cauchy–Riemann Equations

Let $f = u + i v$ be defined on an open set $Ω$ and write $z = x + i y$ . Then $f$ is complex-differentiable at $z_{0} = x_{0} + i y_{0}$ if and only if $u$ and $v$ are real-differentiable at $(x_{0}, y_{0})$ and satisfy the Cauchy–Riemann (CR) equations:

\frac{\partial u}{\partial x} = \frac{\partial v}{\partial y}, \frac{\partial u}{\partial y} = - \frac{\partial v}{\partial x} .

⟳Approach

z_{0}

first horizontally (

h = Δ x

) then vertically (

h = i Δ y

); equating the two expressions for

f^{'} (z_{0})

yields the pair above.

When they hold, the derivative is

f^{'} (z_{0}) = \frac{\partial u}{\partial x} + i \frac{\partial v}{\partial x} = \frac{\partial v}{\partial y} - i \frac{\partial u}{\partial y} .

⟳Take the horizontal-approach limit and substitute the CR relations to rewrite via the

y

-derivatives.

A compact restatement uses the Wirtinger operators:

\frac{\partial}{\partial z ˉ} = \frac{1}{2} (\frac{\partial}{\partial x} + i \frac{\partial}{\partial y}), f analytic ⟺ \frac{\partial f}{\partial z ˉ} = 0.

⟳Treat

z, \overset{z}{ˉ}

as independent coordinates via

x = (z + \overset{z}{ˉ}) /2

y = (z - \overset{z}{ˉ}) / (2 i)

; the CR equations collapse precisely to the vanishing of

\partial_{\overset{z}{ˉ}} f

Consequences of Analyticity

If $f$ is analytic on a domain (open connected set) $Ω$ , then:

$f$ is infinitely complex-differentiable on $Ω$ .
$f$ is analytic in the power-series sense: around each $z_{0} \in Ω$ it equals a convergent Taylor series

f (z) = n = 0 \sum \infty \frac{f ^{(n)} ( z _{0} )}{n !} (z - z_{0})^{n}, ∣ z - z_{0} ∣ < R,

with $R$ at least the distance from $z_{0}$ to $\partial Ω$ .

Both $u$ and $v$ are harmonic: $\nabla^{2} u = \nabla^{2} v = 0$ .

⟳(1)–(2) follow from the Cauchy integral formula by differentiating under the integral; (3) by differentiating the CR equations once more and using equality of mixed partials.

Why It Holds / Derivation

Cauchy–Riemann from Direction-Independence

Step 1. Suppose

f^{'} (z_{0})

exists. Compute the limit in [1] along the real axis,

h = Δ x \to 0

f^{'} (z_{0}) = Δ x \to 0 lim \frac{u ( x _{0} + Δ x , y _{0} ) + i v ( x _{0} + Δ x , y _{0} ) - u - i v}{Δ x} = u_{x} + i v_{x} .

⟳Split numerator into real/imaginary parts; the difference quotients converge to the

x

-partials.

Step 2. Now approach along the imaginary axis,

h = i Δ y \to 0

. Since

1/ (i Δ y) = - i /Δ y

f^{'} (z_{0}) = Δ y \to 0 lim \frac{u ( x _{0} , y _{0} + Δ y ) + i v ( \dots ) - u - i v}{i Δ y} = \frac{1}{i} (u_{y} + i v_{y}) = v_{y} - i u_{x}^{'} -free = v_{y} - i u_{y} .

⟳Multiply numerator and denominator by

- i

; the

y

-partials emerge.

Step 3. Equate the real and imaginary parts of [6] and [7]:

Re : u_{x} = v_{y}

Im : v_{x} = - u_{y}

. These are exactly [2].

■

Real & Imaginary Parts Are Harmonic

Step 1. Assume (by the smoothness consequence) that

u, v \in C^{2}

. Differentiate the first CR equation

u_{x} = v_{y}

with respect to

x

and the second

u_{y} = - v_{x}

with respect to

y

u_{xx} = v_{y x}, u_{y y} = - v_{x y} .

⟳Partial-differentiate each CR identity term by term.

Step 2. Add and use equality of mixed partials

v_{y x} = v_{x y}

u_{xx} + u_{y y} = v_{y x} - v_{x y} = 0.

⟳The mixed second derivatives of

v

cancel by Clairaut's theorem.

An identical argument gives $v_{xx} + v_{y y} = 0$ . Thus both parts solve Laplace's equation. $■$

Bernhard Riemann | 1826–1866

Though Augustin-Louis Cauchy first formulated the differentiability conditions, it was Bernhard Riemann who, in his 1851 doctoral dissertation, recast complex functions geometrically. Riemann viewed an analytic function as a conformal map and introduced Riemann surfaces to tame multivalued functions like $z$ and $lo g z$ . His insight that analyticity is equivalent to the CR system — and his physical interpretation via potential theory and fluid flow — fused complex analysis with mathematical physics. Riemann's name attaches to the equations, to the surfaces, and to the deepest open problem in the field, the Riemann Hypothesis.

Building Intuition

Think of $f$ as a deformation of the plane. The real Jacobian of $(x, y) \mapsto (u, v)$ is the matrix

(u_{x} v_{x} u_{y} v_{y}) .

Imposing the CR equations forces this matrix into the form

(a b - b a) = r (cos θ sin θ - sin θ cos θ),

which is exactly multiplication by the complex number

a + ib = r e^{i θ}

. Geometrically: at each point an analytic map acts as a rotation by $ar g f^{'}$ followed by a scaling by $∣ f^{'} ∣$ — identical in every direction. That is why angles are preserved (conformality) wherever

f^{'} \neq = 0

, and why analytic functions cannot "tear" or "shear" the plane. The rigidity of analyticity — that local data fixes global behavior — is the price the plane pays for this perfect, direction-blind stretching.

An analytic map is locally a rotation-plus-scaling: it preserves angles between curves

Worked Examples

Verifying Analyticity of $f(z)=z^2$

Given:

f (z) = z^{2}

. Find

u, v

, check the CR equations, and compute

f^{'} (z)

. Step 1. Expand with

z = x + i y

z^{2} = (x + i y)^{2} = x^{2} - y^{2} + 2 i x y,

u = x^{2} - y^{2}

and

v = 2 x y

. Step 2. Partial derivatives:

u_{x} = 2 x, u_{y} = - 2 y, v_{x} = 2 y, v_{y} = 2 x .

Step 3. Check CR:

u_{x} = 2 x = v_{y}

✓ and

u_{y} = - 2 y = - v_{x}

✓. The equations hold for all

(x, y)

. Step 4. Derivative via [3]:

f^{'} (z) = u_{x} + i v_{x} = 2 x + i 2 y = 2 (x + i y) = 2 z .

f (z) = z^{2} is entire, with f^{'} (z) = 2 z .

A Function That Is Nowhere Analytic: $f(z)=\bar z$

Given:

f (z) = \overset{z}{ˉ} = x - i y

. Determine where

f

is complex-differentiable. Step 1. Identify parts:

u = x

v = - y

. Step 2. Partials:

u_{x} = 1

u_{y} = 0

v_{x} = 0

v_{y} = - 1

. Step 3. Test first CR equation:

u_{x} = 1

but

v_{y} = - 1

, so

u_{x} \neq = v_{y}

anywhere. Step 4. Via Wirtinger calculus [4]:

\partial f / \partial \overset{z}{ˉ} = \partial (\overset{z}{ˉ}) / \partial \overset{z}{ˉ} = 1 \neq = 0

, confirming non-analyticity.

f (z) = \overset{z}{ˉ} is complex-differentiable at NO point; it is nowhere analytic.

Note: $\overset{z}{ˉ}$ is perfectly smooth as a map $R^{2} \to R^{2}$ — it is a reflection — yet fails complex differentiability because it reverses orientation and cannot be "multiplication by one complex number."

Common Misconceptions

"Satisfying the CR equations at a point implies differentiability there." Not quite — the CR equations alone are necessary but not sufficient for complex differentiability. You also need

u

and

v

to be real-differentiable (e.g. continuous partials). There exist functions satisfying CR at the origin that fail to be complex-differentiable there because the partials are not continuous. The clean sufficient condition: if $u, v$ have continuous first partials in a neighborhood AND satisfy the CR equations, then $f$ is analytic. "Analytic and differentiable mean the same thing as in real calculus." In real analysis,

f (x) = x^{2} sin (1/ x)

is differentiable once but not twice; analyticity (power-series representability) is strictly stronger than differentiability. In complex analysis these notions coincide: complex-differentiable on an open set

\Rightarrow

infinitely differentiable

\Rightarrow

power-series analytic. This collapse has no real-variable analogue.

Consider

f (z) = ⎩ ⎨ ⎧ \frac{z ^{5}}{∣ z ∣ ^{4}}, 0, z \neq = 0 z = 0.

At the origin the Cauchy–Riemann equations are satisfied (the directional derivatives along the axes match), yet the difference quotient

\frac{f ( h ) - 0}{h} = \frac{h ^{5} /∣ h ∣ ^{4}}{h} = \frac{h ^{4}}{∣ h ∣ ^{4}}

depends on

ar g h

(taking different values along

h = t

versus

h = t e^{iπ /4}

). Hence

f^{'} (0)

does not exist despite the CR equations holding pointwise — a vivid demonstration that pointwise CR without differentiability of

u, v

is insufficient.

Connections

Analyticity links to many neighbouring topics:

It underwrites Cauchy's Integral Theorem, which states $\oint_{γ} f d z = 0$ for analytic $f$ on simply connected domains — the foundation of contour integration.
Via Harmonic Functions, the real/imaginary parts solve Laplace's Equation, connecting complex analysis to potential theory and electrostatics.
The power-series consequence ties it to Taylor and Laurent Series, the latter handling isolated singularities.
The geometry of $f^{'} \neq = 0$ is exactly Conformal Mapping, used to solve boundary-value problems by transforming domains.
Failure of analyticity at isolated points launches the Residue Theorem and the classification of Singularities and Poles.

Real-World Applications

2D Electrostatics & Fluid Dynamics. Since $u, v$ are harmonic and conjugate, the complex potential $Φ (z) = ϕ + i ψ$ encodes both the potential $ϕ$ and stream function $ψ$ of an irrotational, incompressible flow or an electrostatic field; flow past a cylinder (Joukowski) is solved exactly this way.
Aerodynamics — Joukowski Airfoils. The conformal map $w = z + 1/ z$ transforms a circle into an airfoil shape, letting engineers compute lift via the Kutta–Joukowski theorem.
Signal Processing & Control Theory. The Laplace and z-transforms produce analytic functions whose poles determine system stability; the Nyquist and Bode criteria are contour-integral arguments.
Heat Conduction. Steady-state temperature distributions in 2D are harmonic functions, computed as real parts of analytic functions, with conformal maps handling complicated geometries.
Quantum Mechanics & Dispersion Relations. The Kramers–Kronig relations relating the real and imaginary parts of a response function follow directly from analyticity in the upper half-plane (causality).
Number Theory. The Riemann zeta function $ζ (s)$ is analytic (except at $s = 1$ ); its analytic continuation and zeros govern the distribution of primes.
Computer Graphics & Imaging. Conformal maps from complex analysis enable texture mapping, fisheye-lens correction, and brain-surface flattening in medical imaging.

Historical Development

1748 — Leonhard Euler relates exponentials and trigonometric functions via $e^{i θ} = cos θ + i sin θ$ , foreshadowing complex function theory.
1797 — Caspar Wessel gives the first geometric (plane) interpretation of complex numbers, later rediscovered by Argand (1806) and Gauss.
1814–1825 — Augustin-Louis Cauchy develops the integral theorem and formulates the differentiability conditions in his work on definite integrals.
1821 — Cauchy's Cours d'analyse establishes rigorous limits, the basis for the modern derivative definition.
1851 — Bernhard Riemann's dissertation reframes analytic functions geometrically, introduces Riemann surfaces, and emphasizes the CR equations and conformality.
1850s–1870s — Karl Weierstrass builds the theory from power series, defining analyticity via convergent series and pioneering analytic continuation.
1859 — Riemann's paper on the prime-counting function introduces $ζ (s)$ as an analytic object, stating the Riemann Hypothesis.
1883 — The term conformal and systematic mapping techniques mature in the hands of Schwarz, Christoffel, and others.
1900s — Édouard Goursat proves the integral theorem assuming only complex differentiability (no continuity of $f^{'}$ ), sharpening the foundations into their modern form.

Subtleties & Limits

The deepest unresolved question intertwined with analyticity is the Riemann Hypothesis (1859): all nontrivial zeros of the analytically continued zeta function $ζ (s)$ lie on the critical line $Re (s) = \frac{1}{2}$ . It remains one of the seven Millennium Prize Problems. More broadly, the theory of analytic continuation raises subtle issues: a function defined by a power series on one disk may extend to a multivalued function on a Riemann surface (e.g. $lo g z$ ), and determining the precise natural boundary beyond which no continuation exists (as for lacunary series like $\sum z^{2^{n}}$ ) is delicate. In several complex variables, the rigidity intensifies — the Hartogs phenomenon shows singularities cannot be isolated — revealing that the one-variable intuition built here is only the beginning.