Holomorphic Functions and Cauchy-Riemann Equations

A function $f : U \to C$ is holomorphic (complex-differentiable on an open set) if and only if it satisfies the Cauchy–Riemann equations $u_{x} = v_{y}$ and $u_{y} = - v_{x}$ together with differentiability of its real and imaginary parts. Key result in words: complex differentiability is enormously stronger than real differentiability — it forces analyticity (local power-series expansion), infinite differentiability, and harmonicity of the real/imaginary parts. Where it applies: complex analysis, conformal mapping, potential theory, fluid flow, electrostatics, signal processing. Who/when: d'Alembert (1752), Euler, Cauchy (1814–1825), Riemann (1851).

A complex function $f$ defined on an open set $U \subseteq C$ is holomorphic at a point $z_{0}$ if the limit $f^{'} (z_{0}) = lim_{h \to 0} \frac{f ( z _{0} + h ) - f ( z _{0} )}{h}$ exists, where $h$ ranges over complex numbers approaching $0$ from every direction. Because $h$ may approach $0$ along any path in the plane, the mere existence of this limit imposes a powerful compatibility constraint on the partial derivatives of the real and imaginary parts of $f$ — precisely the Cauchy–Riemann equations. This single requirement matters because it converts a topological/analytic condition into a rigid algebraic one, and from it cascade the deepest theorems of complex analysis: Cauchy's integral formula, analyticity, the maximum modulus principle, and Liouville's theorem. The surprising fact: a function complex-differentiable merely once on an open set is automatically differentiable infinitely many times and equals its own Taylor series — something with no analogue in real calculus.

The Core Insight

In real calculus, "differentiable" only constrains behavior along the single real line. In the complex plane, differentiability must hold along infinitely many directions of approach simultaneously, and all those directional limits must agree on one common value $f^{'} (z_{0})$ . That coherence condition, written out in terms of $u = Re f$ and $v = Im f$ , is the Cauchy–Riemann equations. Geometrically, $f^{'} (z_{0})$ acts as multiplication by a single complex number — a rotation-plus-scaling — so holomorphic maps are locally conformal (angle-preserving) wherever $f^{'} \neq = 0$ .

Formal Statement

Complex Differentiability and Holomorphy

Let $U \subseteq C$ be open and $f : U \to C$ . We say $f$ is complex-differentiable at $z_{0} \in U$ if the limit

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

exists in $C$ . The function $f$ is holomorphic on $U$ if it is complex-differentiable at every point of $U$ . A function holomorphic on all of $C$ is called entire.

Symbols:

z_{0} = x_{0} + i y_{0}

is the base point;

h = Δ x + i Δ y

is a complex increment;

f^{'} (z_{0}) \in C

is the complex derivative;

U

is an open domain (every point has a neighborhood inside

U

⟳This is the direct transplant of the difference-quotient definition into

C

; the crucial novelty is that division by the complex number

h

requires

h

to approach

0

from all planar directions, not just

\pm

along a line.

The Cauchy–Riemann Equations

Write $f (x + i y) = u (x, y) + i v (x, y)$ with $u, v : U \to R$ real-valued. The Cauchy–Riemann (CR) equations are

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

equivalently written $u_{x} = v_{y}$ and $u_{y} = - v_{x}$ .

Symbols:

u_{x} = \partial u / \partial x

etc. denote first-order partial derivatives. These two real PDEs couple

u

and

v

⟳Compute the limit [1] first along the real direction

h = Δ x \to 0

and then along the imaginary direction

h = i Δ y \to 0

, and equate the two expressions for

f^{'} (z_{0})

Cauchy–Riemann Characterization of Holomorphy

Let $f = u + i v$ be defined on an open set $U$ . Then:

(a) If

f

is complex-differentiable at

z_{0}

, then the partials

u_{x}, u_{y}, v_{x}, v_{y}

exist at

z_{0}

and satisfy the CR equations [2] there, with

f^{'} (z_{0}) = u_{x} + i v_{x} = v_{y} - i u_{y} .

(b) Conversely, if

u, v

are real-differentiable at

z_{0}

(in particular if

u_{x}, u_{y}, v_{x}, v_{y}

exist and are continuous near

z_{0}

) and satisfy the CR equations at

z_{0}

, then

f

is complex-differentiable at

z_{0}

⟳(a) follows from directional limits as in [2]; (b) uses the first-order Taylor (total-differential) expansion of

u

and

v

and substitutes the CR relations to factor the increment as

f^{'} (z_{0}) h + o (∣ h ∣)

Holomorphic ⟹ Harmonic

If $f = u + i v$ is holomorphic on $U$ and (as is automatic) $C^{\infty}$ , then $u$ and $v$ are harmonic:

Δ u = u_{xx} + u_{y y} = 0, Δ v = v_{xx} + v_{y y} = 0,

and $v$ is the harmonic conjugate of $u$ .

⟳Differentiate

u_{x} = v_{y}

x

and

u_{y} = - v_{x}

y

, then add; equality of mixed partials

v_{x y} = v_{y x}

cancels the cross terms.

Why It Holds / Derivation

Deriving the CR Equations from Complex Differentiability

Assume $f^{'} (z_{0})$ exists. We exploit that the limit [1] is independent of the path of approach.

Step 1 — Approach along the real axis. Let

h = Δ x \in R

Δ x \to 0

. Then

\frac{f ( z _{0} + Δ x ) - f ( z _{0} )}{Δ x} = \frac{u ( x _{0} + Δ x , y _{0} ) - u ( x _{0} , y _{0} )}{Δ x} + i \frac{v ( x _{0} + Δ x , y _{0} ) - v ( x _{0} , y _{0} )}{Δ x} .

Taking

Δ x \to 0

gives

f^{'} (z_{0}) = u_{x} (z_{0}) + i v_{x} (z_{0}) .

⟳A real increment leaves the imaginary coordinate fixed, isolating the

x

-partials of

u

and

v

Step 2 — Approach along the imaginary axis. Let

h = i Δ y

Δ y \in R

Δ y \to 0

. Since

1/ (i Δ y) = - i /Δ y

\frac{f ( z _{0} + i Δ y ) - f ( z _{0} )}{i Δ y} = - i [\frac{u ( x _{0} , y _{0} + Δ y ) - u ( x _{0} , y _{0} )}{Δ y} + i \frac{v ( x _{0} , y _{0} + Δ y ) - v ( x _{0} , y _{0} )}{Δ y}] .

Taking

Δ y \to 0

f^{'} (z_{0}) = - i u_{y} (z_{0}) + v_{y} (z_{0}) = v_{y} (z_{0}) - i u_{y} (z_{0}) .

⟳The factor

- i

from

1/ i

swaps and signs the real/imaginary contributions of the

y

-partials.

Step 3 — Equate real and imaginary parts. Because the limit is unique, [5] and [6] must agree:

u_{x} + i v_{x} = v_{y} - i u_{y} .

Matching real parts:

u_{x} = v_{y}

. Matching imaginary parts:

v_{x} = - u_{y}

, i.e.

u_{y} = - v_{x}

. These are exactly the CR equations [2].

■

Converse: CR + Real-Differentiability ⟹ Complex-Differentiability

Assume $u, v$ are real-differentiable at $z_{0}$ with continuous partials satisfying [2]. By the total differential,

u (z_{0} + h) - u (z_{0}) = u_{x} Δ x + u_{y} Δ y + ε_{1} (h), v (z_{0} + h) - v (z_{0}) = v_{x} Δ x + v_{y} Δ y + ε_{2} (h),

where

ε_{j} (h) = o (∣ h ∣)

. Then, with

h = Δ x + i Δ y

Step 1 — Combine.

f (z_{0} + h) - f (z_{0}) = (u_{x} Δ x + u_{y} Δ y) + i (v_{x} Δ x + v_{y} Δ y) + o (∣ h ∣) .

Step 2 — Substitute CR ( $u_{y} = - v_{x}, v_{y} = u_{x}$ ).

= (u_{x} Δ x - v_{x} Δ y) + i (v_{x} Δ x + u_{x} Δ y) + o (∣ h ∣) = (u_{x} + i v_{x}) (Δ x + i Δ y) + o (∣ h ∣) .

Step 3 — Factor. The right side is

(u_{x} + i v_{x}) h + o (∣ h ∣)

, so

\frac{f ( z _{0} + h ) - f ( z _{0} )}{h} = (u_{x} + i v_{x}) + \frac{o ( ∣ h ∣ )}{h} h \to 0 u_{x} + i v_{x} .

Thus

f^{'} (z_{0}) = u_{x} + i v_{x}

exists.

■

⟳The whole trick is that CR is precisely the algebraic condition under which the real-linear map

(Δ x, Δ y) \mapsto (du, dv)

becomes complex-linear, i.e. equals multiplication by a single complex number.

Real-differentiability is needed, not just existence of partials

The mere existence of the four partial derivatives satisfying CR at a single point is not enough for complex differentiability; one needs full real-differentiability (e.g. continuous partials). There exist functions with all partials existing and CR holding at the origin that are not complex-differentiable there. Always verify differentiability of $u, v$ , not just that the partials exist.

Bernhard Riemann | 1826–1866

Though the equations were known to d'Alembert (1752) in fluid dynamics and developed computationally by Euler and Cauchy, it was Bernhard Riemann in his revolutionary 1851 doctoral dissertation who placed them at the geometric heart of function theory. Riemann defined a holomorphic function essentially by the CR system, emphasizing conformality, harmonicity, and the deep link to potential theory, and introduced the notion of a Riemann surface to handle multivaluedness. His geometric, almost physical viewpoint — functions as conformal flows — complemented Cauchy's analytic, integral-based approach and Weierstrass's power-series approach, and together these three perspectives constitute modern complex analysis.

Building Intuition

Think of $f$ as a map deforming the plane. Its real Jacobian at $z_{0}$ is

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

A general real-differentiable map can stretch differently in different directions and shear. But the CR equations force

u_{x} = v_{y}

and

u_{y} = - v_{x}

, so

J = (a b - b a), a = u_{x}, b = v_{x},

which is exactly the matrix of multiplication by the complex number $a + ib$ — a uniform rotation by

ar g (a + ib)

and scaling by

∣ a + ib ∣ = ∣ f^{'} (z_{0}) ∣

. That's why holomorphic maps preserve angles (are conformal) wherever

f^{'} \neq = 0

: there's no shear, no anisotropic stretch, just spin-and-scale. The "infinitely many directions agree" requirement is geometrically the statement "the local map is a similarity transformation."

Conformality: a holomorphic map sends the orthogonal grid through $z_0$ to an orthogonal grid through $f(z_0)$, rotated by $\arg f'$ and scaled by $|f'|$.

Worked Examples

Example 1 — $f(z)=z^2$ is entire; verify CR and find $f'$

Given:

f (z) = z^{2}

. Write

z = x + i y

. Step 1 — Separate into $u, v$ .

z^{2} = (x + i y)^{2} = x^{2} - y^{2} + 2 i x y \Rightarrow u = x^{2} - y^{2}, v = 2 x y .

Step 2 — Compute partials.

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

Step 3 — Check CR equations [2].

u_{x} = 2 x = v_{y} ✓, u_{y} = - 2 y = - v_{x} = - (2 y) ✓ .

Partials are polynomials, hence continuous everywhere, so

f

is holomorphic on all of

C

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

Example 2 — $f(z)=\bar z$ is nowhere holomorphic

Given:

f (z) = \overset{z}{ˉ} = x - i y

. Step 1 — Identify $u, v$ .

u = x, v = - y .

Step 2 — Partials.

u_{x} = 1, u_{y} = 0, v_{x} = 0, v_{y} = - 1.

Step 3 — Test CR equations [2].

u_{x} = 1, v_{y} = - 1 \Rightarrow u_{x} \neq = v_{y} .

CR fails at every point, since

1 \neq = - 1

identically. Step 4 — Confirm via the limit [1]. Along

h = Δ x

: quotient

= \overline{Δ x} /Δ x = 1

. Along

h = i Δ y

: quotient

= \overline{i Δ y} / (i Δ y) = (- i Δ y) / (i Δ y) = - 1

. The two directional limits (

1

and

- 1

) disagree, so

f^{'}

exists nowhere.

f (z) = \overset{z}{ˉ} is real-smooth but nowhere complex-differentiable

Common Misconceptions

CR equations alone are insufficient

Satisfying the CR equations at a point does not guarantee complex differentiability there. You additionally need $u, v$ to be (real-)differentiable, which is guaranteed if the partials exist and are continuous in a neighborhood. CR at an isolated point, without surrounding regularity, can hold while $f^{'}$ fails to exist.

CR holds at a point yet $f$ is not differentiable there

Define

f (z) = ⎩ ⎨ ⎧ \frac{z ˉ ^{2}}{z}, 0, z \neq = 0, z = 0.

At the origin the difference quotients along the axes both give

0

, so the partials of

u

and

v

exist at

0

and one can verify the CR equations hold at the origin. Yet approaching along

z = r e^{i θ}

gives

\frac{f ( z ) - f ( 0 )}{z - 0} = \frac{z ˉ ^{2} / z}{z} = \frac{z ˉ ^{2}}{z ^{2}} = e^{- 4 i θ},

which depends on

θ

. Hence the limit [1] does not exist and

f

is not complex-differentiable at

0

, despite CR holding there. The missing ingredient is real-differentiability of

u, v

at the origin.

Holomorphic is not the same as "smooth" in the real sense

A function can be infinitely real-differentiable (e.g. $f (z) = ∣ z ∣^{2} = z \overset{z}{ˉ}$ ) yet fail to be holomorphic anywhere except possibly isolated points. $∣ z ∣^{2}$ has $u = x^{2} + y^{2}, v = 0$ , so $u_{x} = 2 x, v_{y} = 0$ ; CR forces $x = y = 0$ , meaning $f$ is complex-differentiable only at the origin — and even there it is not holomorphic, since holomorphy requires differentiability on an open set.

Connections

The CR equations are the gateway to the rest of complex analysis. Once holomorphy is established, Cauchy's Integral Theorem shows contour integrals of $f$ over closed curves vanish, and Cauchy's Integral Formula reproduces $f$ inside a contour from its boundary values, instantly yielding analyticity via Power Series and Analyticity. The harmonicity result [4] ties the subject to Harmonic Functions and Potential Theory and Laplace's equation. Conformality links holomorphy to Conformal Mapping and the Riemann Mapping Theorem. The " $\partial / \partial \overset{z}{ˉ} = 0$ " reformulation connects to Wirtinger Derivatives, and the rigidity of holomorphic functions underlies Liouville's Theorem and the Maximum Modulus Principle.

The Wirtinger reformulation

Introducing the operators

\frac{\partial}{\partial z} = \frac{1}{2} (\frac{\partial}{\partial x} - i \frac{\partial}{\partial y}), \frac{\partial}{\partial z ˉ} = \frac{1}{2} (\frac{\partial}{\partial x} + i \frac{\partial}{\partial y}),

the entire CR system collapses to the single elegant equation

\frac{\partial f}{\partial z ˉ} = 0.

Derivation hint for [8]–[9]: Apply $\partial / \partial \overset{z}{ˉ}$ to $f = u + i v$ , expand, and observe that the real and imaginary parts of the result vanish exactly when $u_{x} = v_{y}$ and $u_{y} = - v_{x}$ . Holomorphic functions are precisely those "independent of

\overset{z}{ˉ}

Real-World Applications

Two-dimensional ideal fluid flow. The complex potential $Φ = ϕ + i ψ$ is holomorphic; $ϕ$ is the velocity potential and $ψ$ the stream function, with velocity field $\overline{Φ^{'} (z)}$ . CR equations encode incompressibility and irrotationality; this models airfoil lift via the Joukowski transform.
Electrostatics and steady heat conduction. In source-free regions the potential is harmonic; pairing it with its harmonic conjugate gives a holomorphic complex potential whose level curves are equipotentials and field lines.
Conformal mapping for engineering geometry. Holomorphic bijections map complicated domains to simple ones (e.g. a half-plane), transferring solved boundary-value problems — used in capacitor design, microstrip lines, and grid generation.
Signal processing and the Hilbert transform. Analytic signals are boundary values of functions holomorphic in the upper half-plane; the Hilbert transform pairs a signal with its harmonic conjugate.
Aerodynamics — Joukowski airfoils. A holomorphic map turns a circle (around which flow is trivial) into an airfoil shape, computing lift via the Kutta–Joukowski theorem.
String theory and conformal field theory. Holomorphic and antiholomorphic factorization governs correlation functions on the worldsheet, where conformal invariance is central.
Number theory. Holomorphic modular forms and $L$ -functions (e.g. the Riemann zeta function, holomorphic except for a pole) encode deep arithmetic via their analytic structure.

Historical Development

1752 — Jean le Rond d'Alembert writes down what are essentially the CR equations while studying two-dimensional fluid flow, relating velocity components.
1777 — Leonhard Euler uses the same relations in evaluating real integrals via complex substitution, developing computational facility with $u, v$ pairs.
1814–1825 — Augustin-Louis Cauchy systematically develops the equations within his integral-based theory of complex functions, proving the foundational integral theorem; the equations appear as a necessary condition for path-independence of integrals.
1841 — Cauchy formalizes the role of the equations and contour integration in his published memoirs, establishing the residue calculus.
1851 — Bernhard Riemann, in his doctoral thesis Grundlagen für eine allgemeine Theorie der Functionen einer veränderlichen complexen Größe, makes the equations the geometric definition of a holomorphic function, links them to conformality and Riemann surfaces.
1857 — Riemann applies the theory to Abelian functions and develops potential-theoretic methods (the Dirichlet principle).
1860s–1880s — Karl Weierstrass builds an alternative, rigorous foundation via convergent power series, proving the equivalence of differentiability and analyticity and clarifying logical gaps in the Dirichlet principle.
1900–1902 — Édouard Goursat proves that mere complex differentiability (without assuming continuity of $f^{'}$ ) suffices for Cauchy's theorem (the Cauchy–Goursat theorem), removing earlier regularity hypotheses.
20th century — The Wirtinger calculus and the operator form $\partial f / \partial \overset{z}{ˉ} = 0$ unify the equations with the theory of several complex variables and PDE.

Subtleties & Limits

Several deep refinements and boundaries surround the basic theory. The Looman–Menchoff theorem (1923–1936) shows that if $f$ is continuous on $U$ and its partials exist and satisfy CR everywhere — without assuming continuity of the partials — then $f$ is holomorphic; this sharpens exactly how weak the hypotheses can be, and removing continuity of $f$ breaks the conclusion. In several complex variables, the system $\partial f / \partial \overset{z}{ˉ}_{j} = 0$ becomes an overdetermined PDE (the $\overset{ˉ}{\partial}$ -problem), whose solvability and regularity (Hörmander's $L^{2}$ estimates, the Cauchy–Riemann complex on CR manifolds) remain an active research area with subtle obstructions absent in one variable. The interplay between holomorphy, harmonicity, and the inhomogeneous equation $\overset{ˉ}{\partial} u = f$ underlies modern complex geometry, and questions about boundary behavior, the Dirichlet problem on rough domains, and quasiconformal generalizations (where CR is relaxed to a Beltrami equation) continue to drive analysis.

Note for the applied reader

For physicists: whenever you see a 2D harmonic field with an associated "conjugate" field (potential/stream function, equipotential/field line), you are implicitly using a holomorphic function, and the CR equations are the statement that those two fields are orthogonal and co-determined. The complex derivative $f^{'}$ gives the local field strength and orientation in a single complex number — a major computational economy.