Univalent Functions and Theorems on Mappings

Univalent Functions and the Bieberbach Theorem

Motivation: How Much Can Be “Distorted” by a Univalent Mapping?

The Riemann mapping theorem guarantees the existence of a conformal mapping of any simply connected domain onto the unit disk. But how "large" can the distortion be? The Bieberbach theorem (Bieberbach conjecture, proved by de Branges in 1985) gives an exact answer in terms of the expansion coefficients. This is one of the most famous theorems of classical complex analysis of the 20th century.

The Class S — Univalent Functions

Class S: Functions $f$ that are holomorphic and univalent (injective) in the unit disk $\mathbb{D}$, normalized by $f(0) = 0$, $f'(0) = 1$.

Expansion: $f(z) = z + a_2 z^2 + a_3 z^3 + \ldots$ (under normalization).

The Koebe Function: $k(z) = \frac{z}{(1-z)^2} = z + 2z^2 + 3z^3 + \ldots$ ($a_n = n$). This is the “extremal” function in class $S$: the image $k(\mathbb{D})$ is the whole plane minus the ray $(-\infty, -1/4]$.

The Koebe Theorem

Theorem (Koebe's $1/4$-theorem): The image $f(\mathbb{D})$ for $f \in S$ contains the disk $|w| < 1/4$. The estimate is sharp: for $k(z)$, the image $= \mathbb{C} \setminus (-\infty, -1/4]$.

Corollary (Estimate on the Derivative): For $f \in S$ and $|z| < 1$: $ \frac{1 - |z|}{(1 + |z|)^3} \leq |f'(z)| \leq \frac{1 + |z|}{(1 - |z|)^3}. $

The Bieberbach–de Branges Theorem

Bieberbach Conjecture (1916): For $f \in S$ with $f(z) = z + \sum a_n z^n$: $|a_n| \leq n$ for all $n \geq 2$.

De Branges Theorem (1985): The conjecture is true. Equality is attained only on the Koebe function and its rotations $e^{-i\varphi} k(e^{i\varphi} z)$.

De Branges' proof used the Loewner–Kufarev system and special polynomials—a blend of the theory of quasiconformal mappings, functional analysis, and the theory of special functions.

Quasiconformal Mappings

Generalization of conformal: A $K$-quasiconformal mapping allows "distortion" of angles up to a factor $K$. Formally—it solves the Beltrami equation: $ \frac{\partial \overline{f}}{\partial \overline{z}} = \mu(z) \cdot \frac{\partial f}{\partial z}, \quad \text{where } |\mu(z)| \leq \frac{K-1}{K+1} < 1. $

If $\mu \equiv 0$—conformal. If $\mu \neq 0$—quasiconformal.

Applications: Teichmüller theory (moduli spaces of Riemann surfaces), computer geometry (seamless texture mappings).

Numerical Example

Task: For $f(z) = z + a_2 z^2$, check the condition $|a_2| \leq 2$ from the Bieberbach theorem. For which $a_2$ does the function cease to be univalent in $\mathbb{D}$?

Step 1. $f'(z) = 1 + 2a_2 z$. Critical points: $f'(z_0) = 0 \implies z_0 = -1/(2a_2)$. If $|z_0| \geq 1$, then $f' \neq 0$ on $\mathbb{D}$, so $f$ has no critical points.

Step 2. $|z_0| < 1 \Longleftrightarrow 1/(2|a_2|) < 1 \Longleftrightarrow |a_2| > 1/2$. If $|a_2| > 1/2$, then there is a critical point $z_0$ with $|z_0| < 1$, i.e., inside $\mathbb{D} \implies f$ is not univalent.

Let us clarify: $|a_2| > 2$ violates the class $S$ condition (by Bieberbach). But for $|a_2| > 1/2$, univalence is already violated.

Step 3. Example: $a_2 = 1$ (allowed: $|a_2| = 1 < 2$). $f(z) = z + z^2$. Is $f(z_1) = f(z_2)$ possible? $z_1 + z_1^2 = z_2 + z_2^2 \implies (z_1 - z_2)(1 + z_1 + z_2) = 0$. Roots coincide when $z_1 + z_2 = -1$. If $z_1 = -0.6$, $z_2 = -0.4$: both in $\mathbb{D}$? $|z_1| = 0.6$, $|z_2| = 0.4$—yes. $f(-0.6) = -0.6 + 0.36 = -0.24$, $f(-0.4) = -0.4 + 0.16 = -0.24$—thus, not univalent at $a_2 = 1$!

Conclusion: Bieberbach's $|a_n| \leq n$ is a “coefficient condition” that does not automatically guarantee univalence. Univalence is a global property.

Real-world Application

Computer animation and cartography: quasiconformal mappings are used in algorithms for “smoothed deformation” of polygonal meshes. They provide the “most conformal” possible mapping (minimal angle distortion) under boundary condition constraints.

Additional Aspects

Univalent functions are studied in the class $S$ of holomorphic injections $f: \mathbb{D} \to \mathbb{C}$ with $f(0) = 0$, $f'(0) = 1$. The famous Bieberbach conjecture $|a_n| \leq n$ (where $a_n$ are the Taylor series coefficients of $f$) was formulated in 1916 and proved by de Branges in 1985. The class $S$ is connected to hydrodynamics (jet theory), the theory of stochastic differential equations (SLE — Schramm–Loewner Evolution), and random tree models in mathematical physics. The Koebe distortion theorem asserts that $f(\mathbb{D})$ necessarily contains a disk of radius $1/4$ and provides sharp estimates for $|f'(z)|$ and $|f(z)|$ in terms of $|z|$; the constant $1/4$ is attained for the Koebe function $k(z) = z/(1-z)^2$.

Connection with Other Branches of Mathematics

Univalent functions are closely related to the theory of differential equations via the Loewner equation. In the classical version (Loewner, 1923), one considers the evolution of conformal mappings via a time parameter and a differential equation for the family $f(\cdot, t)$, prescribed by a driver function on the boundary of the disk. The Kufarev generalization leads to a whole theory of subordination chains, and in the late 20th century, Oded Schramm interpreted the stochastic version of this process (SLE) as the solution of a stochastic differential equation, relating univalent functions to two-dimensional conformal invariance in probability theory (percolation cluster boundaries, Ising models).

From a topological viewpoint, univalent functions realize universal coverings: by the Riemann theorem, any simply connected domain on the sphere, except the sphere and plane, is conformally equivalent to the disk, and univalent mappings describe concrete covering maps. The concept of Teichmüller space (Teichmüller, Grötzsch) relies on quasiconformal deformations, and the Bieberbach theorem interacts with variational methods in geometric function theory.

Algebraic aspects manifest themselves in extremal problems: the coefficients of univalent functions form a convex but very complicated compact subset in infinite-dimensional space, and de Branges–type proofs use operator methods (Hilbert spaces, positive kernels, inspired by the works of Hasse, Hopf, Hill). In numerical analysis, the theory of univalent and quasiconformal mappings underlies methods of conformal mapping of domains (works of Gauss, Riesz, Harris, and subsequent authors on numerical solutions of the Dirichlet problem) and mesh partitioning algorithms in computational geometry.

Historical Note and Development of the Idea

The first Bieberbach-type inequality for the coefficient $a_2$ was proved by Bieberbach in 1916 (Mathematische Annalen). At the same time, he formulated the conjecture $|a_n| \leq n$ for all $n$, motivated by problems of conformal geometry and distortion estimates. In the 1920s–1930s, Schiffer, Carl Tietze, and later de la Vallée Poussin, Rogozinsky, Noether proved the conjecture for small $n$, developing variation techniques and Loewner substitution. In 1923, Loewner introduced his parametric method to prove the conjecture for $n = 3$, linking univalent functions with differential equations. Later, Pomerantsev, Goldin, Kufarev, and their school systematized Loewner chains and connected them with hydrodynamics problems (free boundary motion).