MrMineev


Euler Characteristic

Algebraic Topology · ๐Ÿ—“๏ธ October 14, 2026

Given a space \(X\) with a CW structure, let us define the Euler Characteristic of the space as, \[ \chi(X) = \sum_{i} (-1)^i n_i(X) \] where \(n_i(X)\) counts the number of \(i\)-cells of \(X\).

Theorem 1.1

\(\chi(X) = \sum_{i} (-1)^i \rank H_i(X)\)

Proof. Let us start by looking at the cellular chain complex, \[ \ldots \longrightarrow C_{n + 1} \overset{\delta_{n + 1}}{\longrightarrow} C_{n} \overset{\delta_{n}}{\longrightarrow} C_{n - 1} \longrightarrow \ldots \longrightarrow C_0 \longrightarrow 0 \] where \(C_n = H_n(X^n, X^{n - 1})\). Let \(A_n = \ker \delta_n\) and \(B_n = \im \delta_{n}\), then by the fundamental isomorphism theorem and the definition of a homology we have two exact short sequences (since we know that the cellular homology is isomorphic to the singular homology for CW-complexes), \[ \begin{cases} 0 \longrightarrow A_n \longrightarrow C_n \longrightarrow B_{n} \longrightarrow 0 \\ 0 \longrightarrow B_{n + 1} \longrightarrow A_n \longrightarrow H_n \longrightarrow 0 \end{cases} \] now taking the ranks of the two equations we obtain that (since the rank is additive), \[ \rank C_n = \rank A_n + \rank B_n = \rank B_{n + 1} + \rank B_n + \rank H_n \] Thus, \[ \sum_{i} (-1)^i \rank H_n = \sum_{i} (-1)^i \biggl( \rank \im \delta_{n + 1} + \rank \im \delta_n + \rank H_n \biggr) = \sum_{i} (-1)^i \rank C_n = \sum_{i} (-1)^i n_i(X) \] since \(\rank \im \delta_0 = \rank 0 = 0\) and the last equality is due to \(C_k \cong H_k(X^k, X^{k - 1})\) which is isomorphic to the \(k\)-homology of the wedge sum of \(n_k\) \(k\)-spheres and thus a rank of \(n_k\)!

Remark 1.2

Note that the \(\rank C_n\) argument in the end can be more directly written as, \[ \rank H_k(X^k, X^{k - 1}) = \rank H_k \left( \bigvee_{i = 1}^{n_k(X)} S^k \right) = \rank \bigoplus_{i = 1}^{n_k(X)} H_k(S^k) = \rank \mathbb{Z}^{n_k(X)} = n_k(X) \]

The remarkable aspect is that the homology groups do not depend on the CW structure imposed on \(X\), since singular homology is the same as simplicial for CW complexes.


ยง

๐Ÿ’ฌ Comments