1 Definition and terminology

A cluster point is a point that lies arbitrarily close to members of a set or terms of a sequence, in a way that persists no matter how small the neighborhood around the point becomes. The idea captures local concentration: there are always elements distinct from the point itself, or at least infinitely many nearby terms, depending on the context. In analysis, the term is used with closely related meanings, and the exact formulation often depends on whether one is discussing a set, a sequence, or a more general topological space.

1.1 Cluster point of a set

For a set, a cluster point is typically a point such that every neighborhood of it contains at least one element of the set other than the point itself. Equivalently, the set meets every sufficiently small neighborhood around that point in a nontrivial way. This notion does not require the point to belong to the set, although it may.

1.2 Accumulation point

An accumulation point is a point around which the set has elements arbitrarily close, usually meaning that every neighborhood contains infinitely many points of the set or at least one point different from the candidate point. In many texts, accumulation point and cluster point are used interchangeably for sets, though some authors reserve one term for slightly different formulations.

1.3 Limit point

A limit point is another closely related term, especially in topology and real analysis. It often refers to a point that can be approached by points of a set distinct from the point itself. In many elementary settings, limit point and accumulation point coincide, but terminology varies by branch and author.

1.4 Isolated point

An isolated point is a point of a set that is not a cluster point of that set. It has a neighborhood containing no other points of the set. Thus, isolated points stand apart from the accumulation behavior that defines cluster points.

2 Characterizations

Cluster points can be described in several equivalent ways, especially in metric spaces and first-countable topological spaces. These characterizations are useful because they connect geometric intuition with formal definitions and allow different proof techniques.

2.1 Neighborhood-based definition

The neighborhood formulation says that a point is a cluster point if every neighborhood of the point intersects the set in a point other than the point itself, or more generally in infinitely many points. This captures the idea that no matter how tightly one zooms in, the set continues to appear around the point.

2.2 Sequences approaching a point

In metric spaces, a point is often a cluster point of a set if there exists a sequence of distinct points from the set converging to that point. This sequence-based description is especially intuitive, since it translates a local property into the existence of an approximating sequence.

2.3 Topological definition

In general topological spaces, cluster points are defined using neighborhoods rather than distances. A point is a cluster point of a set if every neighborhood of the point intersects the set in a point different from the point itself. This version avoids reliance on metrics and works in broad settings.

2.4 Relationship to closure

Every cluster point of a set belongs to its closure, because closure contains all points that cannot be separated from the set by neighborhoods. However, not every point in the closure must be a cluster point; an isolated point of the set also lies in the closure, even though it is not accumulated upon by other points of the set.

3 Cluster points of sequences

For sequences, cluster points describe values that are approached by subsequences rather than necessarily by the whole sequence. This notion is central to understanding boundedness, convergence, and the behavior of sequences that do not settle at a single limit.

3.1 Subsequence limits

A cluster point of a sequence is a limit of some subsequence. This provides a precise way to identify values that recur in the asymptotic behavior of the sequence, even if the original sequence oscillates or does not converge.

3.2 Convergent subsequences

When a sequence has a convergent subsequence, the limit of that subsequence is a cluster point. This is one of the most important tools in analysis, because it allows one to extract regular behavior from a sequence with potentially irregular overall motion.

3.3 Bounded sequences and compactness

Bounded sequences in Euclidean space often have cluster points, and compactness provides the setting in which such points are guaranteed. The link between boundedness and compactness is especially visible in the Bolzano–Weierstrass theorem, where bounded sequences are shown to possess convergent subsequences.

3.4 Cluster sets

The cluster set of a sequence is the collection of all its cluster points. This set may contain one point, several points, or an entire interval, depending on the sequence. It summarizes the long-term behavior of the sequence more fully than a single limit value can.

4 Properties

Cluster points have structural properties that make them useful in analysis and topology. They interact naturally with limits, neighborhoods, and set-theoretic operations, and their behavior often reflects the geometry of the surrounding space.

4.1 Uniqueness and multiplicity

A set or sequence may have one cluster point, many cluster points, or none at all. A convergent sequence has exactly one cluster point, namely its limit, while an oscillating or dense sequence may have several. The possibility of multiplicity distinguishes cluster behavior from ordinary convergence.

4.2 Dependence on the ambient space

Whether a point is a cluster point can depend on the space in which the set is viewed. A set may behave differently when considered as a subset of a larger topological space, because neighborhoods and closure properties are determined by the ambient structure. This makes context important in precise statements.

4.3 Stability under set operations

Cluster points interact predictably with unions, intersections, and closures, although the exact relationships vary. For example, the cluster points of a union include cluster points coming from either part, while intersections may reduce the number of such points. These relationships are often used in proofs about closed sets and compact sets.

4.4 Relationship with density

A dense set has cluster points throughout the region it densely occupies, though density and cluster points are not identical notions. Density concerns approximation of every point in a space or region, while cluster points focus on points near which the set continues to accumulate. The two ideas are closely connected in many classical examples.

5 Examples

Concrete examples make the distinction between cluster points, isolated points, and ordinary limits easier to see. Simple real-number sets and sequences already show most of the typical behaviors.

5.1 Simple subsets of the real line

The set \(\{1/n : n \in \mathbb{N}\}\) has a cluster point at 0, since the terms approach 0, even though 0 is not in the set. The set \(\{1, 1/2, 1/3, \dots\}\) has no other cluster points in \(\mathbb{R}\). By contrast, a finite set has no cluster points in the usual sense.

5.2 Sequences with one cluster point

The sequence \(a_n = 1/n\) has the single cluster point 0. Although the sequence never equals 0, its terms get arbitrarily close to 0 as \(n\) grows. This is the simplest example of convergence producing a unique cluster point.

5.3 Sequences with multiple cluster points

The sequence \(a_n = (-1)^n\) has two cluster points, \(-1\) and \(1\). Its terms alternate between these values, so subsequences formed by even and odd indices converge to different limits. More elaborate oscillating sequences can have a whole interval of cluster points.

5.4 Sets with no cluster points

A finite subset of a metric space has no cluster points, because each point can be isolated by a small enough neighborhood. Similarly, a discrete set with all points separated from one another may lack cluster points. In such cases, the set is made entirely of isolated points.

Cluster points belong to a broader family of notions describing limiting behavior. Several closely related terms appear in analysis and topology, each emphasizing a slightly different aspect of approximation or convergence.

6.1 Adherence point

An adherence point is a point that belongs to the closure of a set, meaning every neighborhood intersects the set. This is broader than being a cluster point, because the neighborhood may meet the set only at the point itself. Thus, all cluster points are adherence points, but not conversely.

6.2 Limit superior and limit inferior

For sequences of real numbers, the limit superior and limit inferior describe the largest and smallest subsequential limits, respectively, when these exist in the extended real line. They are closely related to cluster points because the cluster set of a bounded sequence lies between these extremal values. In many cases, they provide compact summaries of oscillatory behavior.

6.3 Derived set

The derived set of a set is the collection of all its cluster points. This construction is central in point-set topology, where it helps describe perfect sets, closed sets, and repeated accumulation. Iterating the derived-set operation can reveal deeper structure in a set.

6.4 Compactness and completeness

Compactness often guarantees the existence of cluster points for infinite sequences and sets. Completeness is also relevant, because in complete metric spaces, Cauchy sequences converge, which strongly constrains possible cluster behavior. Together, these ideas form part of the framework in which convergence becomes manageable.

7 Applications in analysis

Cluster points are a basic tool in analysis because they isolate the limiting behavior of sets and sequences. They appear in proofs, theorem statements, and the study of continuity, compactness, and convergence.

7.1 Convergence proofs

When proving convergence, one often shows that every subsequence has the same cluster point, or that the sequence cannot have more than one. This method is useful for establishing uniqueness of limits and for handling sequences defined recursively or indirectly. Cluster points provide a flexible route to limit results.

7.2 Continuity and discontinuity

Cluster points help describe how a function behaves near a point, especially through the values taken by inputs approaching that point. Continuity can be expressed by requiring function values to follow the behavior of nearby points. In contrast, discontinuities often become visible when values near a cluster point fail to approach a single output.

7.3 Bolzano–Weierstrass theorem

The Bolzano–Weierstrass theorem states that every bounded sequence in \(\mathbb{R}^n\) has a convergent subsequence. In other words, boundedness ensures the existence of at least one cluster point. This theorem is one of the foundational results linking compactness, boundedness, and convergence.

7.4 Use in metric and topological spaces

Cluster points are defined in both metric and topological settings, making them adaptable to many branches of analysis. In metric spaces, distances provide intuitive control over neighborhoods and sequences. In topological spaces, the concept remains valid without any notion of distance, which broadens its range of application.