Conformal Mapping

A conformal map is an analytic function looked at with a geometer's eye. Instead of asking what value $f(z)$ takes, we ask how it deforms the plane — and the answer, wherever $f$ is analytic with $f'(z_0)\neq0$ , is startlingly rigid: in the small, every such map is a rotation followed by a uniform stretch. Two curves that cross at an angle map to two image curves that cross at the same angle, with the same sense of turning. That angle preservation is the whole meaning of the word conformal, and it falls straight out of complex differentiability: a single derivative $f'(z_0)$ must serve for approach from every direction, which forces the local Jacobian into the shape of a scaled rotation matrix. The physical payoff comes through harmonic functions. The real and imaginary parts of an analytic function each satisfy Laplace's equation, and a conformal change of variable carries one harmonic function to another. So every planar potential problem — electrostatics, steady heat flow, ideal fluid flow, 2D magnetostatics — can be solved by mapping an awkward boundary onto a straight line or a circle, solving Laplace's equation there for free, and pulling the answer back. Two theorems make this systematic rather than lucky. The Möbius transformations $w=(az+b)/(cz+d)$ are an explicitly computable group that sends circles-and-lines to circles-and-lines and is fixed by three point conditions. The Riemann mapping theorem guarantees that any simply connected proper subdomain of the plane is conformally equivalent to the unit disk. Together they license the physicist's reflex: reduce the hard region to a disk or half-plane, solve, and transport the solution home. Conformality is local, and it fails exactly where $f'=0$ — but even that failure is useful, because it is how a Schwarz–Christoffel map opens out the corners of a polygon.

Key takeaways

A conformal map $w=f(z)$ preserves the angle and orientation between intersecting curves wherever $f$ is analytic and $f'(z_0)\neq0$ .
The local action is $w-w_0\approx f'(z_0)(z-z_0)$ : $|f'(z_0)|$ scales every length and $\arg f'(z_0)$ rotates every direction, so infinitesimal figures map to similar figures.
Because $\operatorname{Re}f$ and $\operatorname{Im}f$ are harmonic, conformal maps transport solutions of Laplace's equation — the engine behind all potential-theory applications.
The Möbius transformations $w=(az+b)/(cz+d)$ with $ad-bc\neq0$ are the prototype family; three point conditions fix one uniquely and they send circles/lines to circles/lines.
The Riemann mapping theorem: every simply connected domain that is not all of $\mathbb{C}$ maps conformally onto the unit disk, making "map the hard region onto a simple one" a universal strategy.
Conformality dies at critical points where $f'(z_0)=0$ : a zero of order $m$ multiplies angles by $m+1$ , which is exactly how corners are engineered.
Key idea: where $f$ is analytic and $f'(z_0)\neq0$ the map is locally $w-w_0\approx f'(z_0)(z-z_0)$ — an exact rotation by $\arg f'(z_0)$ and magnification by $|f'(z_0)|$ — so angles and orientation are preserved and harmonicity is carried between domains.

The core insight

Where $f$ is analytic and $f'(z_0)\neq0$ , the map acts locally as $w-w_0\approx f'(z_0)(z-z_0)$ — an exact rotation by $\arg f'(z_0)$ and magnification by $|f'(z_0)|$ — so angles and orientation are preserved and harmonicity is transported between domains.

Where this lives today

The method is anything but historical. In aerodynamics the Joukowski map $w=z+1/z$ transforms the exactly-solvable flow past a cylinder into the flow past an aerofoil, and the Kutta condition — imposed at the trailing-edge critical point where conformality fails — fixes the circulation and hence the lift, still taught as the analytic backbone of thin-airfoil theory. In microelectronics and MEMS, Schwarz–Christoffel maps compute the capacitance and fringing fields of coplanar strip lines and interdigitated electrodes, where the sharp corners that defeat finite-element meshes are exactly the prevertices the formula handles in closed form. The same maps give the spreading resistance of a constricted contact in heat transfer. And at the frontier, two-dimensional conformal field theory — the engine of string worldsheet dynamics and of critical phenomena in statistical mechanics — rests on the fact that in two dimensions the conformal group is infinite-dimensional, a direct consequence of every analytic function being locally conformal.

The physical picture

Attach an infinitesimal arrow to the point $z_0$ and let $f$ act on it. Differentiability in the complex sense means $f$ has one derivative $f'(z_0)$ that works for approach from every direction, so the image arrow is simply $f'(z_0)$ times the original. Multiplication by a fixed complex number is a rotation by its argument together with a scaling by its modulus — the same rotation and scaling for every arrow at $z_0$ . Two arrows therefore keep the angle between them, and a tiny disk maps to a tiny disk, never to an ellipse. The first figure freezes this local rigidity for the concrete map $f(z)=z^2$ , where the numbers come out clean.

Figure

Local action of f(z) = z² at z₀ = 1 + i: the derivative f′(z₀) = 2 + 2i rotates every direction by 45° and stretches every length by the factor 2√2 ≈ 2.83, so the right angle between the two arrows is carried to a right angle at the image point w₀ = 2i

This matters for physics because harmonic functions ride along for free. Equipotentials and field lines form an orthogonal net, and conformality preserves that orthogonality, so solving Laplace's equation on a simple domain and pulling back drags the whole orthogonal grid into the complicated region. Möbius maps are the sharpest tool for reshaping domains, cleanest on the Riemann sphere, where a line is just a circle through the north pole and "lines and circles map to lines and circles" becomes a single statement. The second figure shows the workhorse of the trade — the Cayley transform, which turns the infinite upper half-plane into the finite unit disk and pins down exactly where the boundary goes.

Figure

The Cayley transform w = (z − i)/(z + i) maps the upper half-plane onto the unit disk: the interior point z = i goes to the centre w = 0, and the real axis wraps once onto the unit circle with z = 0 → −1, z = 1 → −i, z = −1 → i, and z = ∞ → 1, so a boundary of infinite length becomes a circle of circumference 2π

The core equations

A conformal map is nothing more than an analytic function, split into its real and imaginary parts. Writing it this way exposes the two harmonic fields $u$ and $v$ that will do the physical work, and fixes notation for everything below.

w=f(z)=u(x,y)+i\,v(x,y) \qquad (1)

where: $z=x+iy$ — point in the source plane; $w=u+iv$ — image point in the target plane; $u,v$ — real and imaginary parts, each harmonic wherever $f$ is analytic. Reading: one complex equation packages two real fields locked together that each solve Laplace's equation. Analyticity is the Cauchy–Riemann equations. They are the exact condition that the real Jacobian of $(x,y)\mapsto(u,v)$ collapses from four independent entries to the two-parameter shape of a scaled rotation — the algebraic root of conformality.

\frac{\partial u}{\partial x}=\frac{\partial v}{\partial y},\qquad \frac{\partial u}{\partial y}=-\frac{\partial v}{\partial x} \qquad (2)

where: $u_x,v_y$ — the diagonal derivatives, forced equal; $u_y,v_x$ — the off-diagonal derivatives, forced opposite. Reading: these two constraints are precisely what makes $f'$ direction-independent. That direction-independent derivative is best read in polar form, because its modulus and argument are the two geometric quantities the map applies locally.

f'(z_0)=|f'(z_0)|\,e^{i\arg f'(z_0)} \qquad (3)

where: $|f'(z_0)|$ — local magnification, the factor by which lengths near $z_0$ scale; $\arg f'(z_0)$ — local rotation applied uniformly to every direction. Reading: one complex number encodes the entire local geometry of the map. Assembling the local picture over infinitesimal displacements gives the differential statement of conformality, the version you actually manipulate in calculations.

\mathrm{d}w=f'(z)\,\mathrm{d}z \qquad (4)

where: $\mathrm{d}z$ — an infinitesimal displacement in the source plane; $\mathrm{d}w$ — its image, rotated by $\arg f'$ and scaled by $|f'|$ . Reading: displacements transform by complex multiplication, so angles are preserved. The most useful explicit family is the Möbius (bilinear) transformations. They form a group under composition, act on the whole Riemann sphere, and carry the family of circles-and-lines into itself.

w=\frac{az+b}{cz+d},\qquad ad-bc\neq0 \qquad (5)

where: $a,b,c,d$ — complex constants; $ad-bc$ — the nonzero determinant that keeps the map invertible and nonconstant. Reading: with $ad-bc\neq0$ the derivative $f'=(ad-bc)/(cz+d)^2$ never vanishes, so a Möbius map is conformal everywhere except at its pole. What makes Möbius maps practical is that they preserve the cross ratio of any four points — the invariant that lets you write down the unique map sending three points to three targets.

(w_1,w_2;w_3,w_4)=(z_1,z_2;z_3,z_4),\qquad (z_1,z_2;z_3,z_4)=\frac{(z_1-z_3)(z_2-z_4)}{(z_1-z_4)(z_2-z_3)} \qquad (6)

where: the four-point bracket is the cross ratio; equality across the map is the defining property. Reading: three point conditions exhaust the three degrees of freedom of a Möbius map, so it is then unique. The bridge to physics is the transport law for the Laplacian. A conformal change of variable multiplies the Laplacian by the real conformal factor $|f'|^2$ , so harmonicity is preserved wherever the map is conformal, and the disk automorphisms $w=e^{i\theta}(z-a)/(1-\bar a z)$ with $|a|<1$ furnish the natural symmetry group once a domain has been sent to the unit disk.

\nabla_z^2\,(\phi\circ f)=|f'(z)|^2\,\big(\nabla_w^2\phi\big)\circ f \qquad (7)

where: $\phi$ — a field defined in the $w$ -plane; $|f'(z)|^2$ — the positive conformal factor, zero only at critical points. Reading: if $\phi$ is harmonic then $\phi\circ f$ is harmonic, so boundary-value problems transport intact. Finally, when the target domain is a polygon, the required map is given in closed form by the Schwarz–Christoffel formula, which builds the sharp corners by concentrating turning at prescribed prevertices.

\frac{\mathrm{d}w}{\mathrm{d}z}=A\prod_{k=1}^{n}(z-x_k)^{\alpha_k/\pi-1} \qquad (8)

where: $x_k$ — prevertices on the real axis mapping to the polygon vertices; $\alpha_k$ — the interior angle at the $k$ -th vertex; $A$ — a complex constant fixing overall scale and orientation. Reading: each factor injects a fractional-power singularity that bends the boundary through the exterior angle $\pi-\alpha_k$ .

Derivation AWhy analyticity forces a local rotation and uniform scaling

The goal is to show, from the Cauchy–Riemann equations alone, that the real linear map approximating $f$ near a point is not a general $2\times2$ shear-and-stretch but the very special combination of a single rotation with a single isotropic scaling — the algebraic fact from which every geometric consequence (angle preservation, disks-to-disks, orthogonal-net preservation) follows. We start from the real Jacobian, impose Cauchy–Riemann, and recognise the result as complex multiplication.

Step 1. Regard the map as a real transformation $(x,y)\mapsto(u,v)$ . Its linearization at $z_0$ is the Jacobian $J=\begin{pmatrix}u_x&u_y\\ v_x&v_y\end{pmatrix}$ with four in-principle-independent partial derivatives. Why: any smooth planar map looks linear in the small, and a general such linear map distorts angles — so the constraint that saves angles must come from somewhere.

Step 2. Impose the Cauchy–Riemann equations $u_y=-v_x$ and $v_y=u_x$ . The Jacobian collapses to $J=\begin{pmatrix}u_x&-v_x\\ v_x&u_x\end{pmatrix}$ , a matrix of the form $\begin{pmatrix}a&-b\\ b&a\end{pmatrix}$ with $a=u_x$ , $b=v_x$ . Why: the two constraints remove exactly the two degrees of freedom (a shear and an anisotropic stretch) that would have distorted angles, leaving only the two that do not.

Step 3. Factor out $r=\sqrt{a^2+b^2}$ and write $a=r\cos\theta$ , $b=r\sin\theta$ , so $J=r\begin{pmatrix}\cos\theta&-\sin\theta\\ \sin\theta&\cos\theta\end{pmatrix}$ . This is a rotation by $\theta$ scaled by $r$ — identical to multiplication by the complex number $a+ib=f'(z_0)=re^{i\theta}$ . Why: recognizing the CR-Jacobian as a scaled rotation matrix is the same statement as "the derivative is a single complex number", tying the real and complex viewpoints together.

Step 4. Since one rotation and one scaling act on every direction alike, the angle between any two vectors is unchanged and their common orientation is preserved, provided $r=|f'(z_0)|\neq0$ . Why: isotropy is the crux — a map that stretched differently along different axes would distort angles, but $J$ here has a single stretch factor $r$ in all directions. *Note: the CR-Jacobian reduces to $\begin{pmatrix}a&-b\\ b&a\end{pmatrix}$ with $\det J=a^2+b^2=|f'|^2\ge0$ .*

Go deeper · Derivation B · 1 minBuilding the standard upper-half-plane to unit-disk map

Derivation BBuilding the standard upper-half-plane to unit-disk map

The goal is to construct explicitly the Möbius map that carries the upper half-plane onto the unit disk, the single most-used reduction in applied conformal mapping, and to see why it works from the geometry of reflection rather than from brute algebra. The construction rests on one observation: a point and its complex conjugate are mirror images in the real axis, so any real point is equidistant from them, and a ratio of equal distances has modulus one.

Step 1. Pick an interior target point $z_0$ with $\operatorname{Im}z_0>0$ and form $w=e^{i\theta}\dfrac{z-z_0}{z-\bar z_0}$ . Its only pole is at $\bar z_0$ , which lies in the lower half-plane, so the map is analytic throughout the upper half-plane. Why: placing the pole outside the domain guarantees the map stays finite and conformal everywhere we care about.

Step 2. Restrict to the boundary, a real point $z=x$ . Because $z_0$ and $\bar z_0$ are reflections in the real axis, $|x-z_0|=|x-\bar z_0|$ , hence $\bigl|\tfrac{x-z_0}{x-\bar z_0}\bigr|=1$ and $|w|=1$ . So the real axis maps onto the unit circle. Why: this is the entire trick — equal distances to a conjugate pair force unit modulus, sending the boundary of the half-plane to the boundary of the disk.

Step 3. Evaluate at the interior point itself: $z=z_0$ gives $w=0$ . Thus $z_0$ lands at the centre, which certifies that the upper half-plane (the side containing $z_0$ ) maps to the interior $|w|<1$ , not the exterior. Why: fixing where one interior point goes resolves the two-fold ambiguity of which side maps inside.

Step 4. Use the free phase $e^{i\theta}$ to set a desired boundary correspondence. The canonical choice $z_0=i,\ \theta=0$ gives the Cayley transform $w=\dfrac{z-i}{z+i}$ , whose point images $0\mapsto-1$ , $1\mapsto-i$ , $-1\mapsto i$ , $\infty\mapsto1$ wrap the real axis once around the circle. Why: the phase is the last remaining freedom, and choosing it pins the map down to a concrete, tabulated transform. *Note: $w=(z-i)/(z+i)$ maps $\operatorname{Im}z>0$ onto $|w|<1$ , and $|w(x)|=1$ for all real $x$ .*

Go deeper · Derivation C · 2 minSolving a Dirichlet problem by transporting a harmonic function

Derivation CSolving a Dirichlet problem by transporting a harmonic function

The goal is to demonstrate the payoff of the whole subject on a concrete boundary-value problem: find the electrostatic potential in the upper half-plane held at two different constant values on the two halves of the real axis. The strategy is to guess the simplest harmonic function with a jump across the origin — the argument — normalize it to the boundary data, and note it can then be pulled back through any conformal map to solve the same problem on a curved domain.

Step 1. Take $\phi$ harmonic in $\operatorname{Im}z>0$ with $\phi=V_1$ on the negative real axis and $\phi=V_2$ on the positive real axis. The function $\arg z=\operatorname{Im}\log z=\arctan(y/x)$ is harmonic for $y>0$ (it is the imaginary part of the analytic $\log z$ , which has no singularity off the origin). Why: harmonic functions are exactly the imaginary parts of analytic functions, so $\arg z$ is the natural candidate — and it already carries a built-in jump across the branch point at the origin.

Step 2. Fix the constants. As $y\to0^+$ with $x>0$ , $\arg z\to0$ ; with $x<0$ , $\arg z\to\pi$ . So $\phi(x,y)=V_2+\dfrac{V_1-V_2}{\pi}\arg z$ gives $\phi\to V_2$ on the positive axis and $\phi\to V_2+(V_1-V_2)=V_1$ on the negative axis. Why: a linear rescaling of the harmonic building block matches both boundary values while preserving harmonicity, since harmonicity is unaffected by adding a constant or multiplying by one.

Step 3. Read off the field geometry. Equipotentials are rays $\arg z=\text{const}$ from the origin; the orthogonal field lines are concentric semicircles $|z|=\text{const}$ . The two families meet at right angles — the orthogonal net guaranteed by the conformal structure. Why: equipotentials and field lines are the level sets of harmonic conjugates, which analyticity forces to be everywhere orthogonal.

Step 4. Transport. If $w=f(z)$ maps some awkward domain conformally onto this half-plane, then $\phi\circ f$ solves the same two-value boundary problem on that domain, with the ray-and-semicircle net dragged into the new shape. Why: equation (7) says the transported field is still harmonic wherever $f'\neq0$ , so one solved model problem answers a whole family of geometries. *Note: $\nabla^2\arg z=0$ for $y>0$ , and the boundary limits $V_1$ and $V_2$ follow directly from the behaviour of $\arg z$ as $y\to0^+$ .*

Limiting cases

Conformal behaviour is dictated entirely by the local derivative, so the natural limiting cases are the ways $f'$ can behave at a point.

Case	Result	Meaning
Regular point	$f'(z_0)\neq0$	Locally a rotation by $\arg f'(z_0)$ and dilation by $\vert f'(z_0)\vert$ ; angles and orientation preserved — genuinely conformal.
Critical point	$f'(z_0)=0,\ f''(z_0)\neq0$	Behaves like $z\mapsto z^2$ ; angles at $z_0$ are doubled. A zero of order $m$ multiplies angles by $m+1$ .
Isometric scale	$\vert f'(z_0)\vert=1$	Length-preserving rotation; $w=e^{i\alpha}z+b$ is a rigid motion of the plane.
Anticonformal	$w=\overline{f(z)}$	Preserves angle magnitudes but reverses orientation (a reflection); not analytic, so not conformal; at a Möbius pole $z\to-d/c$ the image is $\infty$ , conformal only projectively.

Common trapAnalyticity alone does not make a map conformal

The conformal property needs $f'(z_0)\neq0$ , not merely that $f$ is analytic. At a zero of $f'$ the map ceases to preserve angles: $w=z^2$ is analytic everywhere yet doubles every angle at the origin, because $f'(0)=0$ . More generally a zero of order $m$ multiplies angles by $m+1$ , a feature Schwarz–Christoffel exploits deliberately to build corners. A second, separate trap: the complex conjugation $z\mapsto\bar z$ preserves the size of angles but flips their sense — it is a reflection, anticonformal rather than conformal, and it is not analytic because it violates the Cauchy–Riemann equations. Preserving angle magnitude is necessary but not sufficient; conformality demands orientation be preserved too.

Worked examples

1 · Angle doubling at a critical point, angle preservation away from it

For $f(z)=z^2$ , compare what happens to a pair of directions meeting at $60^\circ$ at the origin with what happens at the regular point $z_0=1+i$ .

Step 1. At the origin, $f'(z)=2z$ gives $f'(0)=0$ , so the linear term dies and the map is governed by the quadratic $w=z^2$ . Writing $z=r e^{i\varphi}$ gives $w=r^2 e^{i2\varphi}$ : the argument is doubled. Two rays leaving the origin at $\varphi_1=20^\circ$ and $\varphi_2=80^\circ$ (a $60^\circ$ opening) map to rays at $40^\circ$ and $160^\circ$ , an opening of $\mathbf{120^\circ}$ . The origin is a critical point and conformality fails there exactly as the $m=1$ rule predicts ( $m+1=2$ ).

Step 2. At $z_0=1+i$ the derivative is $f'(z_0)=2(1+i)=2+2i\neq0$ , so the map is conformal. Its modulus is $|f'(z_0)|=\sqrt{2^2+2^2}=\sqrt8=2\sqrt2\approx\mathbf{2.83}$ and its argument is $\arg(2+2i)=\arctan(2/2)=\mathbf{45^\circ}$ .

Step 3. So near $z_0=1+i$ every direction is rotated by $45^\circ$ and every length stretched by $2.83$ , while any angle between two directions is untouched; the image point is $w_0=(1+i)^2=\mathbf{2i}$ . The same analytic function is angle-doubling at $z=0$ and angle-preserving at $z=1+i$ — conformality is a local, pointwise property tied to $f'\neq0$ .

2 · Verifying the Cayley transform maps the half-plane into the disk

Show that $w=(z-i)/(z+i)$ sends interior points of the upper half-plane strictly inside the unit disk, using $z=2i$ and the centre $z=i$ .

Step 1. Take $z=2i$ . Then $w=\dfrac{2i-i}{2i+i}=\dfrac{i}{3i}=\dfrac{1}{3}$ , so $|w|=\tfrac13<1$ — the point lands strictly inside the disk, as required for an interior point.

Step 2. Take the special point $z=i$ . Then $w=\dfrac{i-i}{i+i}=\dfrac{0}{2i}=0$ — it maps to the centre, confirming that $i$ is the interior point sent to $w=0$ .

Step 3. Check a boundary point $z=1$ (real, on $\partial$ ). Then $w=\dfrac{1-i}{1+i}=\dfrac{(1-i)^2}{(1+i)(1-i)}=\dfrac{-2i}{2}=-i$ , and $|-i|=1$ — a boundary point maps onto the unit circle. Interior maps to interior ( $|w|=\tfrac13$ ), the distinguished point $i$ to the centre, and the boundary to the circle — the map is exactly the half-plane-to-disk reduction.

3 · A two-plate electrostatic potential from a harmonic argument

Two coplanar half-plane electrodes fill the negative and positive real axes; the negative plate is held at $V_1=10\ \mathrm{V}$ and the positive at $V_2=0\ \mathrm{V}$ , with the field confined to $\operatorname{Im}z>0$ . Find the potential at $z=i$ and at $z=1+i$ .

Step 1. From Derivation C the solution is $\phi(x,y)=V_2+\dfrac{V_1-V_2}{\pi}\arg z=\dfrac{10}{\pi}\arg z$ , harmonic for $y>0$ with the right boundary values.

Step 2. At $z=i$ : $\arg z=\pi/2$ , so $\phi=\dfrac{10}{\pi}\cdot\dfrac{\pi}{2}=\mathbf{5\ V}$ — exactly halfway between the plates, as symmetry demands, since $z=i$ bisects the two half-axes.

Step 3. At $z=1+i$ : $\arg z=\arctan(1/1)=\pi/4$ , so $\phi=\dfrac{10}{\pi}\cdot\dfrac{\pi}{4}=\mathbf{2.5\ V}$ . The equipotentials are the rays $\arg z=\text{const}$ , so the potential is linear in the polar angle. $\phi(i)=5\ \mathrm{V}$ and $\phi(1+i)=2.5\ \mathrm{V}$ , read straight off the polar angle — and the same field pulls back through any conformal map to solve the problem on a bent electrode geometry.

Physics of the portrait

Conformal mapping is the meeting point of three centuries of ideas: a computational trick made into geometry by Gauss, into an existence theorem by Riemann, and into an industry by the engineers who used it to design aerofoils and capacitors.

Origins · 1 minBernhard Riemann and how this was discovered

Bernhard Riemann

1826–1866 · geometric function theory and conformal equivalence of domains

Bernhard Riemann turned conformal mapping from a collection of clever substitutions into a structural theory. Complex analysts before him — Euler, Lagrange, and above all Gauss, who in his 1822 prize essay solved the general problem of mapping one surface conformally onto another — knew that analytic functions preserve angles and that particular maps simplify particular regions. What was missing was a guarantee. In his 1851 Göttingen doctoral dissertation Riemann stated what is now called the Riemann mapping theorem: any simply connected planar domain other than the whole plane can be mapped conformally and bijectively onto the open unit disk. This is an existence statement of extraordinary reach — the disk is a universal model for every such domain, so a problem solved on the disk is solved everywhere. Riemann's own argument leaned on the "Dirichlet principle", the assumption that a certain energy integral attains its minimum; Weierstrass exposed the gap, and it took Koebe, Carathéodory, and others in the early twentieth century to supply rigorous proofs. Alongside this, August Ferdinand Möbius had already isolated the bilinear transformations, and Schwarz and Christoffel had written down the polygon formula, so that by Riemann's death in 1866 the two pillars of the subject — an explicit computable group and a sweeping existence theorem — were both in place. The infinite-dimensional richness of two-dimensional conformal symmetry, which Riemann glimpsed, is the same structure that makes two-dimensional conformal field theory so powerful today.

History

1822

— Gauss solves the general problem of conformally mapping one surface onto another

1827

— Möbius isolates the bilinear transformations in his study of projective geometry

1851

— Riemann states the mapping theorem in his Göttingen doctoral dissertation

1869

— Schwarz and Christoffel give the formula mapping a half-plane onto a polygon

1912

— Koebe and Carathéodory supply rigorous proofs of the Riemann mapping theorem

Going further

Texts. Stone & Goldbart, Mathematics for Physics, complex-analysis and conformal-mapping chapters · Arfken & Weber, Mathematical Methods for Physicists, 7th ed., §11 on conformal mapping · Hassani, Mathematical Physics, 2nd ed., chapters on complex analysis and analytic functions · Byron & Fuller, Mathematics of Classical and Quantum Physics, complex-variable chapters · Nakahara, Geometry, Topology and Physics, for the conformal structure behind 2D field theory · Riley, Hobson & Bence, Mathematical Methods for Physics and Engineering, complex-variable chapters · Ahlfors, Complex Analysis, 3rd ed., and Nehari, Conformal Mapping, for the rigorous theory. The key results of this article are: (1) the Cauchy–Riemann Jacobian reduces to the scaled-rotation form $\begin{pmatrix}a&-b\\ b&a\end{pmatrix}$ with $\det=|f'|^2$ ; (2) the Möbius derivative $(ad-bc)/(cz+d)^2$ is nonvanishing when $ad-bc\neq0$ ; (3) the Cayley transform $w=(z-i)/(z+i)$ sends $\operatorname{Im}z>0$ onto $|w|<1$ , with $2i\mapsto\tfrac13$ , $i\mapsto0$ , $1\mapsto-i$ ; (4) $\nabla^2\arg z=0$ for $y>0$ , giving $\phi(i)=5$ V and $\phi(1+i)=2.5$ V for $V_1=10,V_2=0$ ; (5) the Laplacian-transport identity $\nabla_z^2(\phi\circ f)=|f'|^2\nabla_w^2\phi$ . No empirical constants appear; every number is a pure consequence of the maps chosen.

Prerequisites. Analytic Functions and the Cauchy–Riemann Equations · Functions of a Complex Variable and Limits · Harmonic Functions and Conjugates · Elementary Complex Functions. Leads to. Harmonic Functions and Conjugates · Complex Numbers and the Complex Plane · Zeros, Poles, and Isolated Singularities.

Ask the tutor: Give me any planar domain — a wedge, a strip, an eccentric annulus, an aerofoil — and I will walk you through the map that tames it and the harmonic solution it carries home.

What is the difference between an analytic map and a conformal map?

Every conformal map is analytic, but not every analytic map is conformal everywhere. Analyticity is the condition that $f$ has a complex derivative — that it satisfies the Cauchy–Riemann equations. Conformality is the geometric consequence that angles are preserved, and it holds only where the derivative is nonzero. So $f(z)=z^2$ is analytic on the whole plane yet fails to be conformal at the single point $z=0$ , where $f'=0$ and angles double. In practice "conformal map" usually means an analytic function that is also injective on the domain of interest — a genuine one-to-one reshaping with a conformal inverse.

Why does preserving angles make the method so useful in physics?

Because the fields of two-dimensional potential theory are built from harmonic functions, and angle preservation is equivalent to preserving harmonicity. Equipotentials and field lines cross at right angles, and a conformal map keeps that orthogonal net intact while bending it into a new shape. So if you can map a complicated boundary to a simple one — a line or a circle — you can solve Laplace's equation there, where the answer is elementary, and then transport both the potential and its field lines back to the original geometry. The whole labour of the boundary-value problem is offloaded onto finding the right map.

Why are Möbius transformations singled out among all conformal maps?

They are the conformal maps of the entire Riemann sphere onto itself — the only ones globally one-to-one including the point at infinity. That gives them three virtues. They form a group, so they compose and invert cleanly. They send circles-and-lines to circles-and-lines, exactly the family of boundaries that arises in disk and half-plane problems. And they are pinned down by three point conditions through the invariance of the cross ratio, so matching three boundary points writes the map explicitly. No other family combines global invertibility with this much computational convenience.

Does the Riemann mapping theorem tell you how to find the map?

No — and that is the usual frustration. The theorem is a pure existence statement: it promises that a conformal bijection from your simply connected domain to the unit disk exists (and is unique once you fix the image of one interior point and a boundary direction), but it gives no formula. For special domains you have explicit tools — Möbius maps for disks and half-planes, the Joukowski map for aerofoils, Schwarz–Christoffel for polygons — and for the rest you turn to numerics, whose algorithms compute the map to high accuracy precisely because the theorem guarantees there is something to compute.

What actually goes wrong at a critical point where f prime is zero?

The linear approximation vanishes, so the map is no longer locally like multiplication by a nonzero complex number — the leading behaviour is the next nonzero term in the Taylor series. If $f'$ has a zero of order $m$ at $z_0$ , then near $z_0$ the map behaves like $(z-z_0)^{m+1}$ , which multiplies angles by $m+1$ . A right angle becomes a straight angle under $z^2$ ; three rays $120^\circ$ apart get squeezed onto a single line. This is a genuine failure of conformality, but a controllable one: the Schwarz–Christoffel formula deliberately places such singularities at the prevertices to fold the boundary into the sharp corners of a polygon.

Is the complex conjugate map z to z-bar ever useful if it is not conformal?

Yes. It is anticonformal: it preserves the magnitude of every angle but reverses orientation, which is exactly a reflection. Composing an anticonformal map with a conformal one recovers conformality (two reflections make a rotation), and anticonformal maps are the natural language for the Schwarz reflection principle, which extends a harmonic or analytic function across a straight boundary by mirroring it. So conjugation is not a broken conformal map — it is the orientation-reversing half of the fuller symmetry group of angle-magnitude-preserving maps, and it earns its keep whenever a problem has a mirror symmetry.