Local Euler characteristic formula

From HandWiki

In the mathematical field of Galois cohomology, the local Euler characteristic formula is a result due to John Tate that computes the Euler characteristic of the group cohomology of the absolute Galois group GK of a non-archimedean local field K.

Statement

Let K be a non-archimedean local field, let Ks denote a separable closure of K, let GK = Gal(Ks/K) be the absolute Galois group of K, and let Hi(KM) denote the group cohomology of GK with coefficients in M. Since the cohomological dimension of GK is two,[1] Hi(KM) = 0 for i ≥ 3. Therefore, the Euler characteristic only involves the groups with i = 0, 1, 2.

Case of finite modules

Let M be a GK-module of finite order m. The Euler characteristic of M is defined to be[2]

[math]\displaystyle{ \chi(G_K,M)=\frac{\# H^0(K,M)\cdot\# H^2(K,M)}{\# H^1(K,M)} }[/math]

(the ith cohomology groups for i ≥ 3 appear tacitly as their sizes are all one).

Let R denote the ring of integers of K. Tate's result then states that if m is relatively prime to the characteristic of K, then[3]

[math]\displaystyle{ \chi(G_K,M)=\left(\#R/mR\right)^{-1}, }[/math]

i.e. the inverse of the order of the quotient ring R/mR.

Two special cases worth singling out are the following. If the order of M is relatively prime to the characteristic of the residue field of K, then the Euler characteristic is one. If K is a finite extension of the p-adic numbers Qp, and if vp denotes the p-adic valuation, then

[math]\displaystyle{ \chi(G_K,M)=p^{-[K:\mathbf{Q}_p]v_p(m)} }[/math]

where [K:Qp] is the degree of K over Qp.

The Euler characteristic can be rewritten, using local Tate duality, as

[math]\displaystyle{ \chi(G_K,M)=\frac{\# H^0(K,M)\cdot\# H^0(K,M^\prime)}{\# H^1(K,M)} }[/math]

where M is the local Tate dual of M.

Notes

  1. Serre 2002, §II.4.3
  2. The Euler characteristic in a cohomology theory is normally written as an alternating sum of the sizes of the cohomology groups. In this case, the alternating product is more standard.
  3. Milne 2006, Theorem I.2.8

References