Fact-checked by Grok 2 weeks ago

Branch point

In the mathematical field of complex analysis, a branch point of a multi-valued function is a point in the complex plane such that the analytic continuation of the function along any closed path encircling that point results in a value different from the starting value, reflecting the function's multi-valued nature.^[1] This occurs because the function cannot be defined as single-valued and continuous in a full neighborhood of the point without additional structure. Typical examples include the origin for the complex logarithm \log z, where encircling z=0 increases the value by $2\pi i, and for the square root \sqrt{z}, where it causes a sign change.^[2] Branch points are classified into algebraic (finite order, like roots) and logarithmic (infinite order, like logs) types, and they are resolved using branch cuts or Riemann surfaces to define principal branches suitable for computation and analysis. These concepts are crucial for studying the global properties of holomorphic functions and appear in applications from algebraic geometry to physics.^[1]

Fundamentals

Definition

In complex analysis, a branch point is a point in the complex plane where a multi-valued function fails to be single-valued and continuous in any neighborhood of that point, resulting in non-uniqueness of the function's values.^[1] This singularity arises because encircling the branch point along a closed path changes the function's value, preventing a consistent analytic definition without additional structure.^[1] Such points are fundamental to understanding multi-valued functions, where local behavior near the singularity requires distinguishing multiple "branches" of the function.^[3] The concept originated in the 19th century during the study of inverse functions, such as square roots and logarithms, which exhibit multi-valued behavior.^[4] Bernhard Riemann introduced key ideas in his 1851 dissertation, developing Riemann surfaces to model these functions globally, while Karl Weierstrass provided a rigorous foundation through power series expansions and analytic continuation, emphasizing algebraic precision over Riemann's geometric intuition.^[4] Their work, building on earlier contributions like Victor Puiseux's 1850 expansions near singularities, established branch points as essential to the theory of complex functions.^[4] Unlike regular points, where functions are holomorphic, or poles, where singularities are isolated and removable by Laurent series, branch points induce a topological "branching" into multiple sheets, reflecting the function's inherent multi-sheeted structure.^[3] A defining property is that analytic continuation around a branch point yields distinct values based on the path taken, with the transformation of values under closed loops captured by the monodromy group, which describes the permutations of branches.^[1] This path-dependence underscores the need for global tools to resolve the multi-valuedness inherent to such functions.^[3]

Classification

Branch points are classified primarily by their order, which determines the number of sheets in the associated Riemann surface and the nature of the multi-valuedness. Finite-order branch points, also known as algebraic branch points, connect a finite number of sheets h, where encircling the point once permutes the function values among these sheets in a cyclic manner, returning to the original value after h full circuits.^[5] In contrast, infinite-order branch points, such as logarithmic or transcendental ones, involve an infinite number of sheets, where the function values do not return to the initial value after any finite number of encirclings, leading to unbounded multi-valued behavior.^[5]^[6] The monodromy associated with branch points describes the permutation of function values upon analytic continuation around a closed path encircling the point. For finite-order algebraic branch points, this monodromy action generates a cyclic group of order h, reflecting the finite ramification structure.^[5] Logarithmic branch points exhibit additive monodromy, where each encircling adds a constant multiple of $2\pi i to the function value, resulting in no finite cyclic closure.^[6] Branch points differ fundamentally from other types of singularities due to their topological obstruction to defining a single-valued analytic continuation in a neighborhood punctured at the point. Removable singularities allow holomorphic extension by redefining the function at the point, while poles exhibit finite-order Laurent expansions with a principal part of negative powers leading to infinity.^[5] Essential singularities, by contrast, feature Laurent series with infinitely many negative powers, causing the function to take all complex values densely near the point, but without the multi-sheeted structure inherent to branch points.^[6] This multi-valuedness creates a barrier to single-valued holomorphy that is absent in these isolated analytic singularities. Locally near a branch point z_0, the behavior of a multi-valued function f(z) is captured by expansions that reflect the order. For algebraic branch points of finite order h, f(z) admits a Puiseux series of the form

f(z) = \sum_{v=0}^{\infty} A_v (z - z_0)^{v/h},

where the non-integer exponents introduce the fractional branching, and the holomorphic function g(z) = \sum A_v (z - z_0)^{v/h} is single-valued.^[5] For infinite-order logarithmic branch points, the expansion includes a logarithmic term, such as f(z) \sim (z - z_0)^r [\log(z - z_0) + h(z)], where r may be integer or fractional and h(z) is holomorphic, emphasizing the unbounded sheet structure.^[6]

Types

Algebraic Branch Points

Algebraic branch points, also known as finite-order branch points, occur at points z_0 in the complex plane where a multi-valued analytic function exhibits branching behavior characterized by a finite number of sheets in its Riemann surface. These points arise primarily from algebraic functions, such as roots of polynomials, and locally, the function can be expressed in the form (z - z_0)^{p/q} near z_0, where p and q are positive integers with \gcd(p, q) = 1. This fractional power structure implies that the function requires q distinct sheets to become single-valued, as encircling z_0 once permutes the values among these sheets in a cyclic manner of order q. ^[3] The local behavior of an algebraic function around such a branch point is captured by its Puiseux series expansion, a generalization of the Laurent series that incorporates fractional exponents. Specifically, near z_0, the function admits an expansion of the form

f(z) = \sum_{k \geq k_0} a_k (z - z_0)^{k/n},

where n is the branching index (a positive integer equal to the denominator q in reduced form), a_k \in \mathbb{C}, and k_0 is an integer (possibly negative, allowing for poles). This series converges in a punctured disk around z_0 and resolves the multi-valuedness by parameterizing the branches via a uniformizing variable, such as w = (z - z_0)^{1/n}, which maps a neighborhood of w = 0 holomorphically to the function's graph excluding the branch point. ^[3] The branching index n quantifies the degree of ramification at the point, determining the local covering structure: the map from the Riemann surface to the complex plane is n-to-1 near z_0, except at the branch point itself where the preimages coincide, reducing the number of distinct points to one. This ramification index, often denoted \nu_f(p) = n, measures how the function "ramifies" or folds the sheets together, with the total branching number at p given by B(p) = n - 1. In algebraic geometry terms, these points correspond to ramification points of the projection from the curve defined by the algebraic equation to the z-plane, where the derivative vanishes. ^[3] A classic example is the function f(z) = \sqrt{z}, which has an algebraic branch point at z_0 = 0 with branching index n = 2. Locally, f(z) = z^{1/2} expands as a Puiseux series starting with the leading term a_1 z^{1/2}, where higher terms can be zero for this simple case. To see the branching, consider starting at a point z = r > 0 on the positive real axis, where f(r) = \sqrt{r} (principal branch). As one traverses a counterclockwise loop around z = 0 once, the argument of z increases by $2\pi, so f(z) acquires a phase factor of \pi, changing sign to -\sqrt{r}. A second loop restores the original value \sqrt{r}, confirming the order-2 periodicity and the need for two sheets. ^[1]

Logarithmic and Transcendental Branch Points

Logarithmic branch points represent a class of infinite-order singularities in multi-valued analytic functions, distinguished by their additive multi-valuedness arising from a logarithmic term. Near a logarithmic branch point at z_0, the function admits a local expansion of the form f(z) \sim \log(z - z_0) + h(z), where h(z) is a holomorphic function in a neighborhood of z_0.^[1] This form captures the essential behavior, with the principal contribution stemming from the multi-valued logarithm. Encircling the branch point z_0 along a simple closed counterclockwise path results in an increment of $2\pi i to the function value, reflecting the $2\pi change in the argument. This monodromy effect produces an infinite number of sheets on the associated Riemann surface, as repeated encirclements accumulate increments of $2n\pi i for integer n.^[7] The increment under such a loop \gamma around z_0 is rigorously derived from the differential form: \Delta f = \int_\gamma f'(z) \, dz. For the logarithmic case, f'(z) = \frac{1}{z - z_0} + h'(z), where h'(z) is holomorphic at z_0. By the residue theorem, the integral equals $2\pi i times the residue of f'(z) at z_0; the residue of \frac{1}{z - z_0} is 1, while that of h'(z) is 0, yielding \Delta f = 2\pi i.^[7] The monodromy group for a logarithmic branch point is the infinite cyclic group \mathbb{Z}, generated by the $2\pi i shift, which implies a non-compact fundamental group for the punctured domain and manifests as a spiral-like infinite covering in the Riemann surface construction.^[8] Transcendental branch points generalize this infinite-order structure to functions involving essential singularities or irregular multi-valuedness beyond pure logarithms, often arising in non-algebraic contexts. Elliptic integrals provide another key example, such as the complete elliptic integral of the first kind K(k), which exhibits transcendental branch points at the endpoints of the modulus interval (e.g., k = \pm 1), combining algebraic square-root branching with transcendental overall structure and leading to irregular monodromy patterns.^[9] In these cases, the infinite-order monodromy produces coverings that deviate from the simple additive shifts of logarithmic points, often involving denser or more intricate sheet arrangements.

Resolution Methods

Branch Cuts

Branch cuts serve as a practical mechanism to define single-valued branches of multi-valued functions in the complex plane by introducing artificial discontinuities along specified curves. These cuts prevent closed paths that encircle branch points, which would otherwise lead to non-unique values due to the function's multi-valued nature. Typically, a branch cut is constructed as a curve extending from a branch point to infinity or connecting multiple branch points, ensuring that the function remains continuous and analytic in the punctured plane excluding the cut. This approach allows for the selection of a principal branch, where the function is defined consistently within the cut domain.^[10]^[1] The choice of branch cuts is arbitrary, provided they connect the relevant branch points without crossing each other and form a system that avoids enclosing any branch points in loops. Such cuts can take various forms, including straight lines, arcs, or more complex curves, as long as they link the branch points effectively to the boundary at infinity. The specific configuration impacts the domain of the principal branch, restricting analytic continuation to regions not crossed by the cuts, and may influence computational convenience or symmetry in applications. However, the branch points themselves remain invariant regardless of the cut's placement.^[11]^[1] In contour integration, branch cuts modify the evaluation of integrals by imposing a jump discontinuity across the cut, where the difference in function values equals the monodromy increment associated with encircling the enclosed branch points. This discontinuity requires careful deformation of integration paths to avoid or account for the cut, ensuring the integral reflects the chosen branch. As a result, integrals over closed contours that would otherwise encircle branch points become well-defined only when the cuts are respected, altering the residue theorem application in multi-valued settings.^[1]^[11] For a multi-valued function with branch points at z_1, \dots, z_n, a valid cut system must form a tree-like structure connecting these points without creating cycles, thereby preventing any closed path from encircling a subset of branch points independently. This tree configuration ensures the entire system of cuts isolates the multi-valued behavior, allowing a single-valued extension across the complement. The minimal number of such cuts depends on the function's algebraic structure, typically requiring n-1 connections for n finite branch points, plus extensions to infinity if necessary.^[1]^[10]

Riemann Surfaces

Riemann surfaces provide a geometric framework for resolving the multi-valued nature of functions with branch points, constructing a multi-sheeted covering of the complex plane where the function becomes single-valued and holomorphic. For functions with algebraic branch points, such as the roots of a polynomial equation like w^q = z - a, the Riemann surface is formed by taking q copies of the complex plane (sheets) and gluing them along branch cuts emanating from the branch point at z = a. This results in a compact q-sheeted surface if the branch points are finite, ensuring that traversing a closed path around the branch point permutes the sheets according to the branching order.^[3]^[12] In contrast, for functions with logarithmic branch points, such as the principal logarithm \log z, the Riemann surface consists of infinitely many sheets stacked in a helicoid-like spiral structure, where each full encirclement of the origin at z = 0 advances to the next sheet by $2\pi i. This infinite covering resolves the infinite branching by allowing continuous analytic continuation without discontinuity, with sheets connected along a branch cut, typically the negative real axis.^[1]^[13] The Riemann surface X is equipped with a holomorphic projection map \pi: X \to \mathbb{C}, which is a universal covering when X is simply connected, locally homeomorphic except at ramification points corresponding to the branch points in the base plane. These ramification points are where the map \pi has multiplicity greater than one, capturing the branching behavior intrinsically on the surface. The deck transformations of this covering, which are automorphisms of X commuting with \pi, represent the monodromy group, permuting the sheets and encoding the analytic continuation around loops in \mathbb{C} minus the branch points.^[3]^[12] On the Riemann surface X, multi-valued functions like \sqrt{z} or \log z admit analytic continuation to become single-valued holomorphic functions, as the surface topology prevents paths from closing without accounting for sheet changes. This uniformization ensures the function is well-defined and analytic everywhere on X, with the original multi-valuedness arising solely from the projection to the base plane. The monodromy action via deck transformations precisely describes how function values transform under such continuations, resolving ambiguities without artificial cuts.^[13]^[1] Topologically, Riemann surfaces are one-dimensional complex manifolds, often equipped with a hyperbolic metric when the Euler characteristic is negative, contrasting with Euclidean metrics on the plane, though the primary role here is uniformization of the branched function. This structure allows the surface to be a branched cover of \mathbb{C}, with the branch points lifted to regular points on X, facilitating global holomorphy.^[3]^[12]

Applications

Examples in Complex Analysis

One of the simplest examples of a multi-valued function with branch points is the square root function, defined implicitly by w^2 = z. This function possesses branch points at z = 0 and z = \infty, where the two possible values of w coincide.^[14] To render it single-valued and analytic in the complex plane minus a cut, a branch cut is typically placed along the negative real axis from 0 to -\infty. The principal branch is explicitly given by \sqrt{z} = \sqrt{r} \, e^{i \theta / 2}, where z = r e^{i \theta} with r > 0 and \theta \in (-\pi, \pi).^[14] The complex logarithm provides another fundamental illustration of a branch point, specifically of logarithmic type at z = 0. Defined as \log z = \ln |z| + i \Arg z, it is inherently multi-valued, with successive branches differing by multiples of $2\pi i.^[1] A common choice for the branch cut is the ray from 0 along the non-negative real axis to \infty, ensuring continuity in the slit plane; crossing this cut jumps the function to an adjacent branch.^[15] This structure highlights how encircling the origin increments the argument by $2\pi, perpetually switching branches without returning to the original value.^[16] Additional examples include the inverse sine function, \arcsin z, which features algebraic branch points at z = \pm 1. These points arise because the equation w = \sin z maps the complex plane in a way that requires two sheets to resolve the multi-valued inverse, with branch cuts often extending from -1 to -\infty and from 1 to \infty along the real axis.^[17] Similarly, the Gamma function \Gamma(z) is meromorphic with simple poles at the non-positive integers z = 0, -1, -2, \dots, but its logarithm \log \Gamma(z) exhibits logarithmic branch points at these poles due to the function's sign changes encircling each pole. The accumulation of these poles toward infinity creates a dense continuum of such branch points in the neighborhood of z = \infty.^[18] The multi-valuedness of these functions becomes evident through path integration around branch points. For instance, consider \sqrt{z^2 - 1}, which has two finite algebraic branch points at z = \pm 1. Starting on one branch and traversing a closed loop that encircles only one of these points (say, z = 1) while avoiding the other results in the function value switching to the opposite branch upon completion of the path, demonstrating the need for a Riemann surface to globalize the function analytically.^[19] Such visualizations underscore how branch points enforce discontinuities across cuts, preventing single-valued continuation around them.^[1]

Role in Algebraic Geometry

In algebraic geometry, branch points are intimately connected to ramification in morphisms between varieties. For a finite morphism \phi: X \to Y between smooth projective curves over an algebraically closed field, a point p \in X is a ramification point if the differential d\phi_p = 0, meaning the map is not locally an isomorphism at p. The ramification index e_p at p is defined as the multiplicity v_p(\det d\phi), where v_p is the valuation at p, and the image \phi(p) \in Y is termed a branch point. This setup captures how the covering "branches" at those loci, with higher ramification indices indicating greater multiplicity in the fiber.^[20] A key quantitative relation is provided by the Hurwitz formula, which links the topology of the domain and codomain curves through branching data. For a degree-d morphism \phi: X \to Y between compact Riemann surfaces (or equivalently, smooth projective curves), the formula states:

$2g_X - 2 = d(2g_Y - 2) + \sum_{p \in X} (e_p - 1),

where g_X, g_Y are the genera of X and Y, and the sum runs over all ramification points p with e_p > 1. The term \sum (e_p - 1) measures the total branching contribution, showing how ramification increases the genus of the covering curve relative to the base. This formula, originally due to Hurwitz, is fundamental for computing genera in branched covers and extends to more general settings under suitable assumptions.^[20]^[21] Branch points play a crucial role in normalization, the process of desingularizing singular algebraic curves via their integral closure in the function field, which yields a smooth model. For singular curves, the normalization map corresponds to a branched cover resolving singularities, often realized as a Riemann surface in the complex case. A representative example is the Weierstrass equation y^2 = x^3 + a x + b defining an elliptic curve (assuming the discriminant is nonzero), which embeds as a double cover of the projective line \mathbb{P}^1 via the x-projection. Here, the branch points are the three roots of the cubic x^3 + a x + b = 0 (counted with multiplicity) plus the point at infinity, and the normalization ensures the total space is smooth of genus one. This construction highlights how branch points dictate the geometry of the desingularized curve.^[22]^[23] In modern scheme-theoretic terms, the branch locus of a finite morphism of schemes is the support of the discriminant ideal, forming a divisor on the base scheme under purity assumptions, such as when the codomain is normal. This generalizes the curve case to higher-dimensional varieties, where the locus captures ramification hypersurfaces. For unramified aspects in higher dimensions, étale covers—finite étale morphisms with no branching—complement this by probing the étale fundamental group away from the branch locus, enabling the study of Galois representations and local lifting properties.^[24]^[25]

References

[1]
Variations of intronic branchpoint motif: identification and functional ...
Nov 10, 2023 · The branchpoint (BP) motif is an essential intronic element for spliceosomal pre-mRNA splicing. In mammals, its sequence composition, distance to the ...
[2]
Large-scale analysis of branchpoint usage across species and cell ...
Here, we report large-scale mapping of the branchpoints from deep sequencing data in three different species and in the SF3B1 K700E oncogenic mutant background.
[3]
[PDF] Branch Points and Branch Cuts (18.04, MIT). - MIT Mathematics
Oct 11, 1999 · Thus the origin is a branch point of log(z). Definition 1.1 The point z0 is called a branch point | for the complex (multiple) valued function f ...Missing: Weierstrass | Show results with:Weierstrass
[4]
[PDF] A concise course in complex analysis and Riemann surfaces ...
a germ from the Riemann surface representing the branch point (z 6= z1). On the other hand, by definition z = Z1(w) is a chart on N near w1. Hence, this ...
[5]
STUDIES IN THE HISTORY OF COMPLEX FUNCTION THEORY II
Already in the spring of 1856 Dirichlet contrasted Riemann and Weierstrass in a letter to. Kummer in Berlin. Dirichlet remarks here that Riemann had developed ...
[6]
[PDF] Ahlfors, Complex Analysis
Complex Analysis has successfully maintained its place as the standard elementary text on functions of one complex variable. There is, never-.
[7]
[PDF] Introduction to Complex Analysis Michael Taylor
Chapter 2. Going deeper – the Cauchy integral theorem and consequences. 5. The Cauchy integral theorem and the Cauchy integral formula.
[8]
https://mathworld.wolfram.com/LogarithmicBranchPoint.html
[9]
Logarithmic Branch Point -- from Wolfram MathWorld
A branch point whose neighborhood of values wrap around an infinite number of times as their complex arguments are varied.
[10]
(PDF) Branch Cuts and Branch Points for a Selection of Algebraic ...
The 4th column displays the location of the branch cuts as intervals. The 5th column summarizes real and imaginary parts of the jump across a branch cut.
[11]
Branch Cut -- from Wolfram MathWorld
A branch cut is a curve (with ends possibly open, closed, or half-open) in the complex plane across which an analytic multivalued function is discontinuous.
[12]
8.2: Branches - Physics LibreTexts
Apr 30, 2021 · First, we define one or more curve(s) in the complex plane, called branch cuts (the reason for this name will be explained later). Next, we ...<|control11|><|separator|>
[13]
[PDF] Riemann Surfaces - UC Berkeley math
Orientability of Riemann surfaces will follow from our desire to do complex analysis on them; notice that the complex plane carries a natural orientation, in.
[14]
[PDF] Complex Analysis on Riemann Surfaces Contents 1 Introduction
There are no Weierstrass points on a Riemann surface of genus. 1. The branch points of every hyperelliptic surface of genus g ≥ 2 are. Weierstrass points.
[15]
[PDF] 2.4 BRANCHES OF FUNCTIONS
The point z = 0, common to all branch cuts for the multivalued square root function, is called a branch point. The mapping w = fα (z) and its branch cut are ...
[16]
[PDF] The complex logarithm, exponential and power functions
Different values of k correspond to different branches of the multivalued complex logarithm. To better understand the branches of a multivalued function, ...
[17]
[PDF] Complex functions, single and multivalued.
Expand this region in such a way as to leave out z = 0, this leaves a “cut” in the plane around the branch point and extending off to infinity. This cut runs ...<|control11|><|separator|>
[18]
[PDF] The complex inverse trigonometric and hyperbolic functions
These two lines comprise the branch cuts of Arcsinz and Arccosz; each branch cut ends at a branch point located at x = −1 and x = 1, respectively (although the ...
[19]
[PDF] Functions of One Complex Variable - Calvin University
Jan 24, 2008 · The gamma function . ... Note that z = 0 and z = ∞ are branch points; however, they are branch points of infinite order (logarithmic.
[20]
[PDF] Lecture 8: Branches of multi-valued functions - UW Math Department
A branch of a multi-valued function f on E ⊂ C is a function that assigns to each z ∈ E one value from f(z). Page 4. Principal branch of z. 1 n .
[21]
[PDF] FOUNDATIONS OF ALGEBRAIC GEOMETRY CLASSES 39 AND 40
Hence it is supported at a finite number of points. These are called the ramification points (and the images downstairs are called the branch points).
[22]
[PDF] The Riemann-Hurwitz Formula - webspace.science.uu.nl
For a covering ϕ : C → D ramified in P, a point P ∈ C with eP > 1 is called a ramification point and ϕ(P) ∈ D is called a branch point; see 3.2. A p-group will ...
[23]
[PDF] FOUNDATIONS OF ALGEBRAIC GEOMETRY CLASS 47
Apr 22, 2008 · in P1, there exists a unique double cover branched at those 4 points (Class 45, Claim 1.3). Thus elliptic curves correspond to 4 distinct points ...
[24]
[PDF] Lecture Notes Algebraic Geometry III/IV Matt Kerr
The normalization theorem. We state (but do not yet prove) the promised relationship between algebraic curves and Riemann surfaces, and explain how to work it ...
[25]
Section 58.28 (0EA1): Purity of ramification locus—The Stacks project
In this section we discuss the analogue of purity of branch locus for generically finite morphisms. Apparently, this result is due to Gabber.Missing: geometry | Show results with:geometry
[26]
[PDF] Lectures on etale cohomology - James Milne
ifolds in differential geometry, a covering of Riemann surfaces with no branch points in complex analysis, and an unramified extension in algebraic number ...