Complex Analysis

A complex function f is analytic (equivalently, holomorphic) at a point z0 if it is complex-differentiable in some open neighborhood of z0, not merely at z0 itself; f is analytic on an open set if it is analytic at every point of that set. This single requirement is dramatically stronger than real differentiability: an analytic function is automatically infinitely differentiable and equals its own Taylor series in a disk around each point of analyticity. Analyticity is equivalent to satisfying the Cauchy–Riemann equations with continuous partials throughout a neighborhood, and it is the central hypothesis underlying essentially every deep theorem of complex analysis (Cauchy's theorem, the identity theorem, Liouville's theorem).

The complex derivative of f at z0 is defined by the same difference-quotient limit as in single-variable real calculus, f'(z0) = lim_{h→0} [f(z0+h) − f(z0)]/h, except now h ranges over complex numbers approaching 0 from every direction in the plane. Because the limit must agree regardless of the direction of approach, complex differentiability at a point is a far more demanding condition than real differentiability, and it is captured algebraically by the Cauchy–Riemann equations. All the familiar rules — sum, product, quotient, and chain rules — carry over unchanged from real calculus, so polynomials, exponentials, and compositions of differentiable functions are differentiated exactly as expected. The theory built on top of this single definition (analyticity, Cauchy's theorems, residues) is what gives complex analysis its remarkable power and rigidity compared to real analysis.

Cauchy's integral formula expresses the value of an analytic function at any interior point of a simple closed contour C entirely in terms of the values of the function on the boundary C itself: f(a) = (1/2πi) ∮_C f(z)/(z−a) dz, for f analytic on and inside C and a inside C. This remarkable formula shows that a holomorphic function's interior behavior is completely determined by its boundary values — a phenomenon with no real-analysis analogue. Differentiating the formula under the integral sign (justified for analytic functions) produces a formula for every derivative f^(n)(a) as a contour integral, which is what proves that analytic functions are automatically infinitely differentiable.

Cauchy's integral theorem (also called the Cauchy–Goursat theorem) states that if f is analytic throughout a simply connected domain D, then the contour integral of f over any closed contour C lying in D is zero: ∮_C f(z) dz = 0. This is the single most important theorem in complex analysis — it explains why contour integrals of analytic functions are path-independent within a simply connected region and it is the launching point for the Cauchy integral formula, Taylor series, and the residue theorem. The Goursat refinement shows the result holds even without assuming f' is continuous, using only analyticity itself. The theorem fails when the domain has 'holes' punctured by singularities, or when f is not analytic somewhere inside C.

A contour integral generalizes the ordinary definite integral to integration along a curve C in the complex plane rather than along a real interval. If C is parametrized by z(t) = x(t) + iy(t) for t in [a,b], the contour integral of f along C is defined by substituting the parametrization and integrating with respect to t: ∫_C f(z) dz = ∫_a^b f(z(t)) z'(t) dt. Contour integrals depend in general on the path taken, not just its endpoints, unless f is analytic on a simply connected region containing the path (Cauchy's theorem). Closed contours, denoted with the symbol ∮, are especially important because their values are governed by the singularities enclosed, which is the foundation for the residue theorem.

A conformal map is a function that preserves angles locally — if two curves cross at angle θ, their images under the map cross at the same angle θ (including orientation). A key theorem of complex analysis is that any analytic function f with f'(z0) ≠ 0 is automatically conformal at z0: the nonzero derivative acts as a local rotation-and-scaling (multiplication by the complex number f'(z0)), which by definition preserves angles between curves through z0. Conformal maps are central to solving boundary-value problems in physics and engineering, since they can transform a complicated domain (e.g. an airfoil cross-section) into a simple one (e.g. a disk or half-plane) while preserving the harmonic nature of solutions to Laplace's equation.

A Möbius transformation (also called a fractional linear transformation) is a map of the form f(z) = (az+b)/(cz+d), where a, b, c, d are complex constants with ad − bc ≠ 0. These maps are bijections of the extended complex plane (the Riemann sphere, ℂ ∪ {∞}) onto itself, they are conformal everywhere they are defined, and they form a group under composition. Möbius transformations send the set of 'circlines' (circles and straight lines, treating a line as a circle through ∞) to circlines, and they are uniquely determined by their action on any three distinct points. Because of this rigidity and flexibility, Möbius transformations are the standard tool for mapping one classical domain (a disk, a half-plane) conformally onto another.

The Riemann sphere is the extended complex plane ℂ ∪ {∞} realized as an honest compact surface — a sphere — via stereographic projection. Sit the unit sphere in ℝ³ with its south pole touching the origin of the complex plane (or, in the standard convention, centered at the origin with the plane through the equator), and project each point of the sphere from the north pole through to where the line hits the plane. Every point of the sphere except the north pole itself maps to a unique finite complex number, and as points on the sphere approach the north pole, their images run off to infinity in every direction on the plane — so declaring the north pole to correspond to a single point ∞ compactifies the plane into a sphere with no boundary and no distinguished points. This turns ℂ ∪ {∞} into a genuine compact Riemann surface on which Möbius transformations act as the full group of holomorphic automorphisms, and on which every rational function becomes a well-defined map of the sphere to itself, poles included.

The Schwarz lemma is a rigidity statement about holomorphic self-maps of the unit disk that fix the origin: if f is holomorphic on the open unit disk D, maps D into D, and f(0) = 0, then |f(z)| ≤ |z| for every z in D, and also |f'(0)| ≤ 1. Moreover, if equality holds anywhere in either inequality (short of the trivial case), f must be a rotation, f(z) = e^(iθ)z for some real θ. The proof is a slick one-line application of the maximum modulus principle to the auxiliary function g(z) = f(z)/z (made removable-singularity-free at 0 by g(0) = f'(0)). Despite its short proof, the Schwarz lemma is enormously consequential: it is the seed of the Schwarz–Pick lemma (which extends it to maps not fixing the origin, producing the hyperbolic metric on the disk), and it underlies rigidity and uniqueness results throughout geometric function theory, including the classification of automorphisms of the disk.

A Laurent series generalizes the Taylor series to functions that are analytic on an annulus (a ring-shaped region) rather than a full disk, allowing negative powers of (z − z0) in addition to nonnegative ones. If f is analytic in the annulus r < |z − z0| < R, it has a unique expansion f(z) = Σ_{n=−∞}^{∞} a_n (z − z0)^n valid throughout the annulus, where the coefficients are given by contour integrals over any circle inside the annulus. The part with negative powers, Σ_{n=1}^{∞} a_{-n}(z-z0)^{-n}, is called the principal part and encodes the singular behavior of f at z0; the coefficient a_{-1} is the residue, central to the residue theorem. Laurent series are the essential tool for classifying isolated singularities as removable, poles, or essential.

An isolated singularity of f is a point z0 where f fails to be analytic but f is analytic on some punctured neighborhood 0 < |z − z0| < r. The Laurent series of f about z0 classifies the singularity into exactly one of three types: removable (no negative-power terms; f extends analytically to z0), a pole of order m (the principal part terminates, with a_{-m} ≠ 0 the lowest nonzero negative coefficient), or essential (infinitely many nonzero negative-power terms). Poles are the most common and useful type, since near a pole of order m, f(z) behaves like c/(z−z0)^m and |f(z)| → ∞ as z → z0. Essential singularities are far wilder: by the Casorati–Weierstrass theorem, f takes values arbitrarily close to every complex number in any neighborhood of an essential singularity.

The residue theorem generalizes Cauchy's integral formula to functions with several isolated singularities inside a closed contour: if f is analytic inside and on a simple closed contour C except for finitely many isolated singularities z1, ..., zn inside C, then ∮_C f(z) dz = 2πi times the sum of the residues of f at those singularities. This turns the calculus problem of evaluating a contour integral into the algebra problem of computing residues (Laurent coefficients a_{-1}), and for a pole of order m, the residue can be computed by an explicit derivative formula without expanding the full Laurent series. The theorem is the single most powerful computational tool in complex analysis, and it is routinely used to evaluate real definite integrals that are difficult or impossible by elementary real methods.

The argument principle counts the zeros and poles of a meromorphic function enclosed by a contour, using only a single contour integral of f'/f. If f is meromorphic inside and on a positively oriented simple closed contour C, with no zeros or poles on C itself, then (1/2πi)∮_C f'(z)/f(z) dz equals Z − P, where Z is the number of zeros and P the number of poles inside C, each counted with multiplicity. The name comes from an equivalent description: this integral equals (1/2π) times the total change in the argument (angle) of f(z) as z traverses C once, i.e. it counts how many times the image curve f(C) winds around the origin. This turns questions like 'how many roots does this polynomial have in this region' into a winding-number computation, and it underlies Rouché's theorem, a workhorse for locating roots without solving equations directly.

The Mandelbrot set is the set of complex numbers c for which the sequence z→z²+c, starting at z=0, stays bounded forever rather than escaping to infinity. Despite that simple rule, plotting which points belong to the set produces one of the most intricate and famous images in mathematics — infinitely detailed at every zoom level.

The complex plane (also called the Argand plane or z-plane) is the plane formed by identifying each complex number z = x + iy with the point (x, y) in a Cartesian coordinate system. The horizontal axis, called the real axis, carries the real numbers; the vertical axis, called the imaginary axis, carries the purely imaginary numbers. This identification turns complex-number arithmetic into plane geometry: addition becomes vector addition, and multiplication becomes a combined scaling and rotation. The complex plane is the natural setting for complex analysis, where functions of a complex variable are studied via their geometric action on this plane.

The Cauchy–Riemann equations are a pair of partial differential equations that characterize when a complex function is differentiable in the complex sense (holomorphic). Writing a complex function f(z) = f(x+iy) = u(x,y) + i v(x,y) in terms of its real part u and imaginary part v, f is complex-differentiable at a point exactly when u and v satisfy ∂u/∂x = ∂v/∂y and ∂u/∂y = −∂v/∂x there (together with u, v having continuous partial derivatives, which gives complex differentiability, not just satisfaction of the equations). These equations are far more restrictive than ordinary real differentiability: they force a two-variable real function pair (u, v) to behave, as a whole, like a single smooth complex function — the root reason complex analysis is so much more rigid than real analysis.

A complex power series is an infinite series of the form Σ aₙ(z − z₀)ⁿ, where the coefficients aₙ and center z₀ are complex numbers and z ranges over the complex plane. Just as with real power series, every complex power series has a radius of convergence R ≥ 0 (possibly infinite): the series converges absolutely for every z with |z − z₀| < R and diverges for every z with |z − z₀| > R, while the boundary circle |z − z₀| = R is a genuine circle of ambiguity where convergence must be checked point by point. What makes the complex setting so much richer than the real line is geometric: the region of guaranteed convergence is an actual disk in the plane, not just an interval, and a function defined by a convergent power series is automatically holomorphic inside that disk, with a derivative obtained by differentiating the series term by term. This is the concrete computational bridge between formal Taylor series and the abstract notion of an analytic function.

A real-valued function u(x, y) of two real variables is harmonic on a region if it has continuous second partial derivatives there and satisfies Laplace's equation, ∂²u/∂x² + ∂²u/∂y² = 0. Harmonic functions arise naturally throughout physics as steady-state temperature distributions, electrostatic potentials, and velocity potentials of incompressible irrotational fluid flow. The deep link to complex analysis is that the real and imaginary parts of any holomorphic function are automatically harmonic — a direct consequence of the Cauchy–Riemann equations — and conversely, every harmonic function on a simply connected domain is the real part of some holomorphic function, called its harmonic conjugate. This correspondence lets the entire machinery of complex analysis (conformal mapping, the maximum modulus principle, mean value properties) be brought to bear on two-dimensional potential problems.

The maximum modulus principle says that if f is holomorphic and non-constant on a connected open domain, then |f| cannot attain a local maximum at any interior point of that domain — every local max of |f| must occur on the boundary. Equivalently, on a compact domain, the maximum of |f| over the closed region is attained on the boundary (never strictly inside, unless f is constant). This is a direct consequence of the mean value property of holomorphic functions: since f(z₀) is the average of f over any small circle around z₀, |f(z₀)| cannot exceed the maximum of |f| on that circle unless f is constant near z₀. The principle is one of the sharpest rigidity results in complex analysis — it rules out an entire class of behaviors (interior 'bumps' in |f|) that are common for smooth real functions, and it underlies uniqueness results, growth estimates like the Schwarz lemma, and many applications in engineering stability analysis.

Functions like the complex logarithm and the square root are not single-valued: every nonzero z has infinitely many logarithms (differing by multiples of 2πi) and every nonzero z has two square roots. To turn such a rule into an honest analytic function, you must choose a branch — a consistent, continuous selection of one value at each point on a restricted domain. The obstruction to doing this on all of ℂ\{0} is topological: walking a small loop around the origin changes arg(z) by 2π, so log(z) = ln|z| + i·arg(z) cannot return to its starting value after a full loop unless you forbid such loops. A branch cut is a curve (often a ray or segment) removed from the domain specifically to block these loops, making arg(z) — and hence log(z) and z^(1/2) — single-valued and holomorphic on what remains. The endpoint(s) of the cut, where the multivaluedness is unavoidable no matter how you cut, are called branch points.

The open mapping theorem states that a non-constant holomorphic function on a connected open domain sends open sets to open sets: if f is holomorphic and non-constant on a domain Ω, then for every open U ⊆ Ω, f(U) is open in ℂ. This is sharply different from smooth real functions, where a non-constant smooth map (like f(x) = x², sending an open interval around 0 to a half-open interval) can easily fail to be open. The theorem follows from a local counting argument: near any point z₀, a non-constant holomorphic f behaves like f(z) = f(z₀) + c(z − z₀)^k + higher order terms for some k ≥ 1 (its local degree), and this k-to-1 behavior near z₀ means f actually covers a whole neighborhood of f(z₀), not just the single point or a boundary sliver. The open mapping theorem is the direct route to a one-line proof of the maximum modulus principle (an open image can't have |f| maximized at an interior point, since nearby image points would have to include ones of strictly larger modulus), and it underlies the inverse function theorem for holomorphic maps at points where f' ≠ 0.

Liouville's theorem states that every bounded entire function is constant: if f is holomorphic on all of ℂ and there is a constant M with |f(z)| ≤ M for every z, then f must be constant. This is a striking rigidity statement with no real-analysis counterpart — plenty of bounded, smooth, non-constant real functions exist (like sin x), but complex differentiability everywhere is so restrictive that boundedness alone forces triviality. The theorem follows quickly from the generalized Cauchy integral formula applied to f' on circles of growing radius: the derivative estimate shows |f'(z)| shrinks to 0 everywhere, so f' ≡ 0 and f is constant. Liouville's theorem is best known as the key step in the fundamental theorem of algebra: assuming a non-constant polynomial had no root leads to 1/p(z) being a bounded entire function, which Liouville's theorem rules out unless it's constant — a contradiction.

Liouville's theorem says a bounded entire function must be constant, so every non-constant entire function is unbounded — but 'unbounded' comes in wildly different flavors, from the mild growth of a polynomial to the explosive growth of e^(e^z). The order of growth ρ of an entire function f is a single real number (or ∞) that measures how fast max|z|=r |f(z)| grows compared to exponentials of powers of r: ρ is the infimum of exponents α such that |f(z)| ≤ exp(|z|^α) eventually. Polynomials have order 0, e^z and sin(z), cos(z) have order 1, and e^(z^n) has order n; functions like e^(e^z) have infinite order. This single number turns out to control deep structural facts, most famously via the Hadamard factorization theorem: an entire function of finite order ρ can be written as a product over its zeros times e^(polynomial of degree ≤ ρ), generalizing how a polynomial factors over its roots — this is the tool that made possible, for instance, Hadamard's product formula for the Riemann zeta function's completed form and proofs related to the prime number theorem.

A meromorphic function on an open set U ⊂ ℂ is one that is holomorphic on U except at a discrete set of poles. At each pole, the function tends to infinity but only like 1/(z−a)^n for some finite n (the order of the pole). Meromorphic functions can be thought of as 'holomorphic functions that are allowed to blow up nicely.' The set of meromorphic functions on a domain forms a field under pointwise addition and multiplication. Key examples include rational functions, tan z, cot z, sec z, and the Gamma function.

The Riemann Mapping Theorem is one of the most powerful results in complex analysis: every simply connected proper open subset of the complex plane is conformally equivalent to the open unit disk. This means that topologically the only distinction between such domains is whether they are simply connected, and conformally there is essentially a unique biholomorphic map from each to the disk (once a basepoint and a direction are normalized). The theorem guarantees existence but the proof — using normal families (Montel's theorem) — is non-constructive; finding explicit maps for specific domains is a major subject in its own right.

The Weierstrass Factorization Theorem generalizes the fact that a polynomial is determined (up to a constant) by its zeros: every entire function can be written as a (possibly infinite) product over its zeros, multiplied by a nonvanishing entire factor e^{g(z)}. The key tool is the Weierstrass elementary factor E_p(z) = (1−z) exp(z + z²/2 + … + z^p/p), which is designed to have exactly one zero at z=1 while converging absolutely in infinite products. This theorem is the foundation for representing classical functions like sin(πz) and the gamma function as infinite products.

Analytic continuation is the process of extending the domain of a holomorphic function beyond its original region of definition. The key insight is that a holomorphic function is completely determined by its values on any open set, so there is at most one way to continue it. This principle underlies the global theory of functions like the Riemann zeta function and the gamma function.

Elliptic functions are doubly periodic meromorphic functions on the complex plane. They arise naturally when inverting elliptic integrals and play a central role in number theory, algebraic geometry, and mathematical physics. The classical examples are the Weierstrass ℘-function and the Jacobi elliptic functions sn, cn, dn.

Modular forms are holomorphic functions on the upper half-plane that transform in a specific way under the action of the modular group \(\text{SL}_2(\mathbb{Z})\). They sit at the crossroads of complex analysis, number theory, and algebraic geometry, encoding arithmetic information in their Fourier coefficients. Fermat's Last Theorem was proved using the modularity theorem for elliptic curves.

A Riemann surface is a one-dimensional complex manifold — a surface on which complex analysis can be performed globally, removing the ambiguities of multi-valued functions. The complex logarithm, square root, and algebraic functions all become single-valued on appropriately constructed Riemann surfaces.

The theory of functions of several complex variables (SCV) extends one-variable complex analysis to \(\mathbb{C}^n\) for \(n \geq 2\). Many results from one variable generalize, but significant new phenomena appear: the Hartogs extension theorem shows that isolated singularities cannot exist in \(n \geq 2\), and domains of holomorphy replace the arbitrary domains of one-variable theory.