Mapping cone of a chain map and the distinguished triangle

Intuition Beginner

A map between complexes compares two systems of equations and cancellations. The mapping cone turns that comparison into a single new complex.

If the map is a perfect cohomological comparison, the cone has no cohomology. If the map misses something, the cone records the missing information.

This makes the cone a homological error detector. Instead of asking only whether a map is good, we build an object that measures how good it is.

The cone is also the algebraic source of triangles. In derived categories, triangles replace short exact sequences as the basic exactness pattern.

Visual Beginner

The cone packages the source, target, and map into one complex.

Worked example Beginner

Suppose a map of complexes includes a subcomplex into a larger complex. The cone behaves like the quotient information, but with a shift built in.

If the inclusion captures all cohomology, the cone has zero cohomology. If the larger complex has new cohomology not seen by the subcomplex, the cone detects it.

What this tells us: cones are the algebraic mechanism behind long exact sequences and distinguished triangles.

Check your understanding Beginner

Formal definition Intermediate+

Let $$ f^\bullet\to D^\bullet $$ be a cochain map in an abelian category $\mathcal{A}$. The mapping cone is the cochain complex $\mathrm{Cone}(f)$ with $$ \operatorname{Cone}(f)^n=C^{n+1}\oplus D^n $$ and differential $$ d_{\operatorname{Cone}(f)}^n= \begin{pmatrix} -d_C^{n+1} & 0\ f^{n+1} & d_D^n \end{pmatrix}. $$ Thus $$ d_{\operatorname{Cone}(f)}(c,d)=(-d_C c,\ f(c)+d_D d). $$

There are natural maps of complexes $$ D^\bullet\to \operatorname{Cone}(f)\to C^\bullet[1], $$ given in degree $n$ by inclusion into the second summand and projection onto the first summand. Together with $f$, these form the standard cone triangle $$ C^\bullet\xrightarrow{f}D^\bullet\to \operatorname{Cone}(f)\to C^\bullet[1]. $$

This triangle is the algebraic analogue of a Puppe sequence, but it is not the same object as the topological mapping cone of a continuous map. The analogy is structural: both constructions turn a map into a cofiber-like object.

Counterexamples to common slips

• The cone is not only the quotient of $D$ by $C$; it includes a shifted copy of $C$.
• The sign in the cone differential is part of the convention that makes the differential square to zero.
• The cone of a quasi-isomorphism is acyclic, but not every acyclic cone is contractible.

Key theorem with proof Intermediate+

Theorem (cone long exact sequence). A cochain map $f:C^{\bullet }\to D^{\bullet }$ gives a long exact sequence $$ \cdots\to H^n(C)\xrightarrow{H^n(f)}H^n(D)\to H^n(\operatorname{Cone}(f))\to H^{n+1}(C)\to\cdots. $$

Proof. By definition, there is a short exact sequence of complexes $$ 0\to D^\bullet\to \operatorname{Cone}(f)\to C^\bullet[1]\to 0. $$ The first map includes $D^n$ as the second summand of $\mathrm{Cone}(f{)}^n$, and the second map projects to $C^{n+1}=C[1{]}^n$.

Apply the long exact cohomology sequence from 01.02.30. Since $$ H^n(C[1])\cong H^{n+1}(C), $$ the resulting sequence has the displayed form.

The connecting morphism identifies with the map induced by $f$, up to the sign dictated by the chosen shift convention. With the convention of 01.02.31, this gives the standard cone triangle and its associated long exact sequence. $\square$

Bridge. The mapping cone is where chain homotopy 01.02.31 becomes triangular structure. The next major step is to axiomatize these cone triangles as distinguished triangles in 04.03.10.

Exercises Intermediate+

Advanced results Master

The mapping cone is the algebraic cofiber of a chain map. It is the object that turns a two-term comparison $$ C^\bullet\xrightarrow{f}D^\bullet $$ into a three-term exactness pattern.

In the homotopy category $K(\mathcal{A})$, the standard cone triangle is the prototype of a distinguished triangle. Triangulated categories axiomatize the formal behavior of these triangles: rotation, functoriality up to controlled choices, and the octahedral axiom for composable maps.

The cone also characterizes quasi-isomorphisms. A map $f$ is a quasi-isomorphism if and only if $\mathrm{Cone}(f)$ is acyclic. This turns the problem of testing a map into the problem of testing a single complex.

In geometry, cone triangles encode many familiar exact sequences. A short exact sequence of complexes can be represented by a cone triangle in the derived category. Long exact sequences in cohomology are then shadows of triangles under the cohomological functor $H^n$.

The analogy with topology is useful but limited. The topological mapping cone produces a cofiber sequence of spaces; applying singular chains gives an algebraic cone up to homotopy-equivalent conventions. The algebraic construction is the chain-level model that survives in derived categories.

Synthesis. The mapping cone is the bridge from complexes to triangulated categories. It measures the failure of a map to be a quasi-isomorphism, produces the standard long exact sequence, and supplies the model triangle from which Verdier's axioms are abstracted.

Full proof set Master

Proposition 1 (the cone differential squares to zero). The block matrix differential on $\mathrm{Cone}(f)$ satisfies $d_{\mathrm{Cone}(f)}^{n+1}{d}_{\mathrm{Cone}(f)}^n=0$.

Proof. Compute on a pair $(c,d)\in C^{n+1}\oplus D^n$: $$ d(c,d)=(-d_Cc,\ f(c)+d_Dd). $$ Applying $d$ again gives $$ d^2(c,d)=(d_C^2c,\ -f(d_Cc)+d_Df(c)+d_D^2d). $$ The first and third terms vanish because $C$ and $D$ are complexes. The middle terms cancel because $f$ is a chain map, so $d_{D}f=f{d}_C$. Hence $d^2=0$. $\square$

Proposition 2 (cone acyclicity detects quasi-isomorphisms). A cochain map $f:C^{\bullet }\to D^{\bullet }$ is a quasi-isomorphism if and only if $\mathrm{Cone}(f)$ is acyclic.

Proof. Use the long exact cone sequence. If all maps $H^n(f)$ are isomorphisms, exactness forces $H^n(\mathrm{Cone}(f))=0$ for every $n$. Conversely, if the cone cohomology vanishes, exactness shows each $H^n(f)$ has zero kernel and full cokernel, hence is an isomorphism. $\square$

Proposition 3 (homotopic maps have homotopy-equivalent cones). If $f\simeq g:C^{\bullet }\to D^{\bullet }$, then $\mathrm{Cone}(f)$ and $\mathrm{Cone}(g)$ are isomorphic in $K(\mathcal{A})$.

Proof. Let $s$ be a homotopy with $f-g=d_{D}s+s{d}_C$. Define a degreewise map from $\mathrm{Cone}(f)$ to $\mathrm{Cone}(g)$ by $$ (c,d)\mapsto (c,d+s(c)). $$ The homotopy equation is exactly the condition that this map commute with the cone differentials. Its inverse is given, up to the matching sign convention, by subtracting the same homotopy term. Therefore the cones are isomorphic in the homotopy category. $\square$

Connections Master

• Chain complexes 01.02.30. The cone is a new complex built from a map of complexes.

• Chain homotopy 01.02.31. Cone behavior is invariant under chain homotopy in the homotopy category.

• Triangulated category 04.03.10. Distinguished triangles axiomatize the formal behavior of cone triangles.

• Derived category 04.03.11. The cone detects quasi-isomorphisms, the maps inverted in the derived category.

• Donaldson-Floer surgery triangle 03.07.25. Floer exact triangles use mapping-cone logic at chain level.

Historical & philosophical context Master

The mapping cone grew from the topological cofiber construction and from the algebraic need to package a map of complexes into a new complex. In topology, mapping cones organize long exact sequences of relative homology; in homological algebra, they organize long exact sequences of cohomology.

Gelfand-Manin and Weibel present cones as the point where ordinary complex manipulation begins to look triangulated ^{[Gelfand-Manin] [Weibel]}. Verdier made this structure axiomatic in the theory of triangulated and derived categories ^[Verdier].

Philosophically, the cone changes the status of a map. Instead of treating a map as external data between two objects, it turns the map into an object whose cohomology measures the map's defect.

Bibliography Master

@book{GelfandManinMappingCone,
  author = {Gelfand, Sergei I. and Manin, Yuri I.},
  title = {Methods of Homological Algebra},
  publisher = {Springer},
  year = {1996}
}

@book{Weibel1994MappingCone,
  author = {Weibel, Charles A.},
  title = {An Introduction to Homological Algebra},
  publisher = {Cambridge University Press},
  year = {1994}
}

@book{Verdier1996MappingCone,
  author = {Verdier, Jean-Louis},
  title = {Des categories derivees des categories abeliennes},
  series = {Asterisque},
  volume = {239},
  publisher = {Societe Mathematique de France},
  year = {1996}
}

@book{CartanEilenberg1956MappingCone,
  author = {Cartan, Henri and Eilenberg, Samuel},
  title = {Homological Algebra},
  publisher = {Princeton University Press},
  year = {1956}
}