ConceptComplete

Mayer-Vietoris and Computations

The Mayer-Vietoris sequence is the primary computational tool for de Rham cohomology, allowing one to compute the cohomology of complicated spaces by decomposing them into simpler pieces.


The Mayer-Vietoris Sequence

Theorem5.2Mayer-Vietoris sequence

Let M=U∪VM = U \cup V where U,VU, V are open subsets. There is a long exact sequence in de Rham cohomology:

⋯→Hk(M)→rHk(U)⊕Hk(V)→sHk(U∩V)→δHk+1(M)→⋯\cdots \to H^k(M) \xrightarrow{r} H^k(U) \oplus H^k(V) \xrightarrow{s} H^k(U \cap V) \xrightarrow{\delta} H^{k+1}(M) \to \cdots

where r[ω]=([ω∣U],[ω∣V])r[\omega] = ([\omega|_U], [\omega|_V]) is the restriction map, s([α],[β])=[α∣U∩V−β∣U∩V]s([\alpha], [\beta]) = [\alpha|_{U \cap V} - \beta|_{U \cap V}], and δ\delta is the connecting homomorphism.

DefinitionConnecting homomorphism

The connecting homomorphism δ:Hk(U∩V)→Hk+1(M)\delta: H^k(U \cap V) \to H^{k+1}(M) is constructed as follows. Given a closed kk-form ω\omega on U∩VU \cap V, choose a partition of unity {ρU,ρV}\{\rho_U, \rho_V\} subordinate to {U,V}\{U, V\}. Define ηV=d(ρUω)\eta_V = d(\rho_U \omega) on VV (extended by zero) and ηU=−d(ρVω)\eta_U = -d(\rho_V \omega) on UU. These agree on U∩VU \cap V and patch to a closed (k+1)(k+1)-form on MM, giving δ[ω]\delta[\omega].


Computations

ExampleCohomology of the torus $T^2$

Write T2=U∪VT^2 = U \cup V where UU and VV are open cylinders overlapping in two disjoint annuli, so U∩V≃S1⊔S1U \cap V \simeq S^1 \sqcup S^1. The Mayer-Vietoris sequence gives:

H0(T2)≅R,H1(T2)≅R2,H2(T2)≅R.H^0(T^2) \cong \mathbb{R}, \quad H^1(T^2) \cong \mathbb{R}^2, \quad H^2(T^2) \cong \mathbb{R}.

The generators of H1H^1 correspond to the two independent loops (meridian and longitude), and the generator of H2H^2 is the area form.

ExampleCohomology of $S^n$ by induction

Cover SnS^n by U=Sn∖{N}U = S^n \setminus \{N\} and V=Sn∖{S}V = S^n \setminus \{S\} (removing north and south poles). Both are contractible (≅Rn\cong \mathbb{R}^n) and U∩V≃Sn−1U \cap V \simeq S^{n-1}. The Mayer-Vietoris sequence gives:

For k≥2k \geq 2: Hk(Sn)≅Hk−1(Sn−1)H^k(S^n) \cong H^{k-1}(S^{n-1}). By induction: Hk(Sn)=RH^k(S^n) = \mathbb{R} if k=0k = 0 or nn, and 00 otherwise.


The Kunneth Formula

Theorem5.3Kunneth formula

For smooth manifolds MM and NN with H∗(M)H^*(M) finite-dimensional,

HdRk(M×N)≅⨁p+q=kHdRp(M)⊗HdRq(N).H^k_{\mathrm{dR}}(M \times N) \cong \bigoplus_{p+q=k} H^p_{\mathrm{dR}}(M) \otimes H^q_{\mathrm{dR}}(N).
ExampleKunneth for the torus

Tn=(S1)nT^n = (S^1)^n. By Kunneth, Hk(Tn)≅(nk)RH^k(T^n) \cong \binom{n}{k} \mathbb{R}, so the Betti numbers are binomial coefficients: bk=(nk)b_k = \binom{n}{k}. The total dimension is ∑bk=2n\sum b_k = 2^n.

RemarkRing structure

The wedge product ∧:Ωp(M)×Ωq(M)→Ωp+q(M)\wedge: \Omega^p(M) \times \Omega^q(M) \to \Omega^{p+q}(M) descends to cohomology, giving HdR∗(M)H^*_{\mathrm{dR}}(M) the structure of a graded-commutative algebra. This cup product structure carries more information than the individual cohomology groups alone.