1 Definition and basic structure
A chain complex is a graded algebraic object equipped with boundary maps that connect neighboring degrees. Its defining feature is that two successive boundary maps compose to zero. This simple condition makes chain complexes a foundational tool in homological algebra and algebraic topology, where they organize data by degree and encode how pieces fit together.
1.1 Sequence of modules or abelian groups
A chain complex is typically written as a sequence \[ \cdots \to C_{n+1} \to C_n \to C_{n-1} \to \cdots \] where each \(C_n\) is an abelian group, module, or vector space. The index \(n\) records the degree of each term. In many settings, the groups are concentrated in finitely many degrees, but infinite sequences are common as well.
The choice of category depends on the application. In topology, the terms often come from free abelian groups generated by geometric objects such as simplices or cells. In algebra, the terms may be modules over a ring. Over a field, chain complexes of vector spaces are especially tractable.
1.2 Boundary maps
The maps between consecutive terms are called boundary maps or differentials. They measure how an object in one degree relates to its “boundary” in the next lower degree. Notationally, one usually writes \[ d_n : C_n \to C_{n-1}. \] These maps are part of the structure of the complex, and the entire collection of groups together with the differentials determines the chain complex.
Boundary maps are linear in the appropriate sense for the setting. Their algebraic form often reflects geometric incidence data, such as which faces belong to a simplex or which cells attach along a boundary.
1.3 The condition d² = 0
The essential axiom of a chain complex is \[ d_{n-1} \circ d_n = 0 \] for every \(n\). This is commonly summarized as \(d^2 = 0\). It expresses the principle that the boundary of a boundary is zero.
This condition has important consequences. It guarantees that images of boundary maps lie inside kernels of the next maps, which makes it possible to define homology groups. Without \(d^2 = 0\), the standard theory of homological invariants would not work.
1.4 Grading conventions
Chain complexes are graded by degree, and conventions differ slightly across subjects. In the chain convention, degrees usually decrease under the differential. In the cochain convention, degrees increase; this is addressed separately in the section on dual notions.
Some authors index chain complexes homologically, while others adopt cohomological grading. Care is needed when comparing formulas, especially for signs in constructions such as tensor products, shifts, and mapping cones. Despite these differences, the underlying idea remains the same: neighboring degrees are linked by a differential whose square is zero.
2 Examples
Examples help show how abstract definitions arise from concrete situations. Chain complexes appear naturally in algebra, geometry, and combinatorics, often encoding combinatorial structure in a linear form.
2.1 Finite chain complexes
A finite chain complex has only finitely many nonzero terms. For instance, \[ 0 \to C_2 \to C_1 \to C_0 \to 0 \] is a finite complex of length three. Such complexes are common in computations and in algebraic models of finite geometric objects.
Finite complexes are often used as test cases because their homology can be computed directly by linear algebra. They also serve as building blocks in more elaborate resolutions and derived constructions.
2.2 Chain complexes from topological spaces
A topological space can often be studied through a chain complex derived from a decomposition into pieces. The resulting algebraic object reflects the arrangement of those pieces and the way they attach to one another.
For example, a triangulated or CW-decomposed space yields a chain complex whose homology captures topological features such as connected components, holes, and voids. This translates geometric information into a form that is often easier to analyze algebraically.
2.3 Simplicial chain complexes
In simplicial topology, one associates to each simplicial degree a free abelian group generated by oriented simplices. The differential sends each simplex to an alternating sum of its faces. This sign pattern is chosen so that successive differentials compose to zero.
Simplicial chain complexes provide a direct combinatorial route to homology. They are especially useful for spaces presented by simplicial complexes, where the geometry is encoded by vertices, edges, triangles, and higher-dimensional simplices.
2.4 Cellular chain complexes
For a CW complex, the cellular chain complex is built from the cells in each dimension. Each group is generated by the cells of a fixed dimension, and the differential records how higher-dimensional cells attach to lower-dimensional ones.
Cellular chain complexes are often smaller than simplicial ones, so they can make calculations more efficient. When available, they give a concise representation of a space’s homological structure while preserving the essential invariants.
3 Homology of a chain complex
Homology is the principal invariant extracted from a chain complex. It measures the extent to which the complex fails to be exact, identifying algebraic classes that are closed under the differential but not themselves boundaries.
3.1 Cycles and boundaries
An element \(x \in C_n\) is called a cycle if \(d_n(x) = 0\). It is called a boundary if there exists \(y \in C_{n+1}\) such that \(d_{n+1}(y) = x\). Because \(d^2 = 0\), every boundary is automatically a cycle.
This leads to a natural containment: \[ \operatorname{im}(d_{n+1}) \subseteq \ker(d_n). \] The gap between these two subgroups is what homology measures.
3.2 Homology groups
The \(n\)th homology group of a chain complex is defined by \[ H_n(C) = \ker(d_n) / \operatorname{im}(d_{n+1}). \] Its elements are cycles modulo boundaries. When the homology group is trivial, every cycle is already a boundary.
Homology groups can be abelian groups, vector spaces, or modules, depending on the setting. They often reveal subtle structural information that is invisible from the chain groups alone.
3.3 Interpretation of homology
Homology provides a way to classify the algebraic “obstructions” present in a complex. In topology, it detects features such as disconnectedness and higher-dimensional holes. In algebra, it measures the failure of algebraic sequences to be exact.
The interpretation depends on context, but the basic idea is stable: homology identifies what survives after boundaries are factored out. This makes it a powerful summary of the information encoded by a chain complex.
4 Morphisms of chain complexes
Chain complexes are not only objects of study; they also admit maps between them. These morphisms preserve the graded structure and commute with the differentials in a suitable way.
4.1 Chain maps
A chain map is a family of homomorphisms between corresponding degrees of two chain complexes that commutes with the differentials. If \(f_n : C_n \to D_n\), then the compatibility condition is \[ d^D_n \circ f_n = f_{n-1} \circ d^C_n. \] This means the map respects the boundary structure.
Chain maps induce homomorphisms on homology. As a result, they are the natural notion of morphism in the category of chain complexes.
4.2 Chain homotopies
Two chain maps can be related by a chain homotopy, which is an algebraic analogue of continuous deformation. A chain homotopy provides intermediate data showing that the two maps differ by a controlled boundary term.
Chain homotopy is weaker than equality but still strong enough to imply that the maps induce the same map on homology. Thus homology is invariant under chain homotopy.
4.3 Homotopy equivalence
A chain homotopy equivalence is a pair of chain maps between two complexes whose compositions are homotopic to the identity maps. Complexes related in this way are considered equivalent from the viewpoint of homological invariants.
Homotopy equivalence is central because it identifies chain complexes that have the same essential algebraic content. Many constructions in homological algebra aim to replace a complex by a homotopy equivalent one that is easier to work with.
5 Exactness and special classes
Special classes of chain complexes are defined by how closely they approximate exactness. These notions help isolate complexes with particularly simple or useful homological behavior.
5.1 Exact complexes
A chain complex is exact at degree \(n\) if \[ \operatorname{im}(d_{n+1}) = \ker(d_n). \] A complex is exact if it is exact in every degree. Exact complexes have zero homology in all degrees.
Exactness means there are no nontrivial homology classes. Such complexes often serve as algebraic models of “perfectly solved” sequences, where every cycle is accounted for by a boundary.
5.2 Acyclic complexes
An acyclic complex is one whose homology groups vanish. In many contexts, this is the same as being exact, though conventions about augmentation may vary in special cases.
Acyclicity is useful because it indicates that the complex carries no homological information beyond what is already built into its structure. Acyclic complexes appear frequently in resolutions and comparison arguments.
5.3 Contractible complexes
A contractible complex is homotopy equivalent to the zero complex. Such a complex is not only acyclic but also homologically trivial in a stronger sense.
Contractible complexes often arise as auxiliary objects in proofs. They can be removed without changing homotopy type, making them useful for simplifying chain-level arguments.
6 Constructions on chain complexes
Chain complexes can be combined and modified in many standard ways. These constructions are important both for theory and for applications, since they allow one to build new complexes from old ones.
6.1 Shift and suspension
A shift reindexes the degrees of a chain complex, moving each term up or down by a fixed amount. This is a basic operation in derived and homotopical settings. Depending on convention, the shift may be denoted by \(C[1]\), \(C[-1]\), or by a suspension-like symbol.
Shifts are especially important for sign conventions and for defining exact triangles in derived categories. They alter degree while preserving the overall pattern of the complex.
6.2 Direct sums and products
Given two chain complexes, one can form their direct sum degree by degree. This produces a new complex whose homology is often easier to understand from the summands. In suitable categories, degreewise products are also available.
Direct sums are useful for assembling complexes from components, while products are more common in infinite constructions. Both preserve the chain structure when the differentials are defined componentwise.
6.3 Tensor products
The tensor product of chain complexes combines algebraic data from two sources. Its differential is defined using a graded sign rule to ensure that the square of the differential remains zero.
Tensor products play a major role in multilinear algebra and topology. They allow the construction of complexes representing combined spaces or modules, and they interact naturally with homology through various comparison theorems.
6.4 Mapping cones
The mapping cone of a chain map is a new complex that measures how far the map is from being a quasi-isomorphism. It packages the source, target, and the map itself into a single object.
Mapping cones are central in derived and triangulated settings. They are often used to define exact triangles and to analyze morphisms by converting them into complexes whose homology reflects the effect of the map.
7 Chain complexes in algebra
Beyond topology, chain complexes are indispensable in algebra. They organize presentations of modules and provide a framework for defining and computing derived functors.
7.1 Free resolutions
A free resolution of a module is an exact chain complex of free modules ending at the module in question. It replaces a potentially complicated object with a more manageable one built from free pieces.
Free resolutions are especially useful because free modules are easy to handle algebraically. They allow homological invariants to be computed by applying functors to a resolution rather than directly to the original module.
7.2 Projective resolutions
A projective resolution is similar to a free resolution, but the modules are projective rather than free. Projective modules retain enough lifting properties to support homological computations while offering greater flexibility.
These resolutions are fundamental in defining derived functors such as Tor. They are also useful when free resolutions are unavailable or inconvenient.
7.3 Injective resolutions
Dually, an injective resolution embeds a module into an exact complex of injective modules. Injective objects are well suited for cohomological constructions because of their extension properties.
Injective resolutions are used to define right-derived functors such as Ext. They provide the cochain-side analogue of projective resolutions and are central in homological methods in algebra.
7.4 Derived functors
Derived functors arise when a functor is extended to chain complexes in order to measure failure of exactness. They capture higher-order information not visible at the level of ordinary functorial images.
Common examples include Tor and Ext. These are defined using resolutions and homology or cohomology of associated complexes, making chain complexes the basic computational framework of derived algebra.
8 Dual notions
Chain complexes have a dual version in which the differential raises degree rather than lowering it. This cochain perspective is especially natural in cohomology theories.
8.1 Cochain complexes
A cochain complex has groups \[ \cdots \to C^{n-1} \to C^n \to C^{n+1} \to \cdots \] with differentials \(d^n : C^n \to C^{n+1}\) satisfying \(d^{n+1} \circ d^n = 0\). The formal structure mirrors that of a chain complex, but the grading direction is reversed.
Cochain complexes are common in algebraic topology and algebraic geometry. They are often more natural when one studies functions, forms, or dual algebraic objects.
8.2 Cohomology
Cohomology is defined from a cochain complex in the same way homology is defined from a chain complex: \[ H^n(C) = \ker(d^n) / \operatorname{im}(d^{n-1}). \] It measures cocycles modulo coboundaries.
Cohomology frequently carries extra algebraic structure, such as cup products, that make it richer than homology in some settings. It complements homology by detecting dual aspects of the same underlying object.
8.3 Relation between chains and cochains
Chains and cochains are related by dualization in suitable categories. For finite-dimensional vector spaces, taking linear duals converts chain complexes into cochain complexes and vice versa.
The relationship is not always perfect in infinite settings, where duality can behave subtly. Even so, the two viewpoints are closely linked and often illuminate each other in geometry and algebra.
9 Applications
Chain complexes appear in many branches of mathematics because they encode structure in a form suited to calculation and abstraction. Their applications range from proving theoretical results to computing invariants of specific objects.
9.1 Homological algebra
Homological algebra is the natural home of chain complexes. It studies exactness, resolutions, derived functors, and the behavior of complexes under various operations.
In this field, chain complexes are not merely auxiliary tools; they are central objects. Many major constructions are formulated directly in terms of maps between complexes and the homology they produce.
9.2 Algebraic topology
Algebraic topology uses chain complexes to convert spaces into algebraic data. The resulting homology groups distinguish spaces that may be difficult to compare by geometric means alone.
Through simplicial and cellular chain complexes, one can compute invariants of spaces built from combinatorial pieces. This approach has made homology one of the standard invariants of topology.
9.3 Computation of invariants
Chain complexes are widely used to compute invariants in both algebra and topology. By organizing a problem into degrees and differentials, they make it possible to apply systematic algebraic methods.
The invariants obtained from homology and cohomology can classify objects up to suitable equivalence, detect hidden structure, and support comparisons across different mathematical settings. In this way, chain complexes function as a bridge between raw data and conceptual understanding.