On the vanishing of group cohomology (Q910862)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On the vanishing of group cohomology
scientific article

    Statements

    On the vanishing of group cohomology (English)
    0 references
    0 references
    0 references
    0 references
    1990
    0 references
    Let \(H^ n(G,M)\) denote the \(n\)-th Tate cohomology group of the finite group \(G\) with coefficients in the \(kG\)-module \(M\), where \(k\) is a field of characteristic \(p\). The paper deals with the question of when the groups \(H^ n(G,M)\) are trivial. Methods of homological algebra, modular representation theory, and advanced structure theory of finite groups are combined to obtain far-reaching results. There is no doubt that this paper is an important contribution to the cohomology theory of finite groups. The first result of the paper is that if \(H^ n(G,M)\neq 0\) for some \(n\) then \(H^ n(G,M)\neq 0\) for infinitely many values of \(n\). This settles a long standing conjecture. In the course of the proof of this result the following new and interesting concept is introduced. A module \(M\) is called a trivial homology (TH) module if there is a finite complex \(C\) \[ 0\to C_ r\to...\to C_ 2\to C_ 1\to C_ 0\to 0 \] of projective \(kG\)-modules \(C_ i\) with \(H_ 0(C)=M\) and \(H_ i(C)\) a trivial \(kG\)-module for \(i=1,2,...,r\). This concept also plays a crucial role in the second question the authors adress: Which groups \(G\) have the property that every non-projective module in the principal block \(B_ 0(kG)\) has non-trivial cohomology. It is well known (and easy to see) that modules outside the principal block have no cohomology, and also that there are groups (for example \(M_{11})\) admitting non-projective modules in the principal block with no homology. Here it is shown that if every module in \(B_ 0(kG)\) is a TH-module, then every non-projective module in \(B_ 0(kG)\) has non-trivial homology. Examples, where this can be shown to be true, are given: \(G\) a group with dihedral Sylow 2-subgroup, \(p=2\); \(A_ 6\), \(p=3\); \(PSL_ 2(8)\), \(p=2\). Next the authors show how to construct a non-projective module (with trivial source and vertex \(P\)) in \(B_ 0(kG)\) with trivial cohomology, if \(P\) is a \(p\)-subgroup whose centralizer is not \(p\)-nilpotent. Thus, if every non-projective module in \(B_ 0(kG)\) has non-trivial cohomology, then the centralizer of every element of order \(p\) in \(G\) is \(p\)-nilpotent. It is conjectured that the converse is also true, and for \(p\neq 2\) this is proved by a deep and involved argument. The authors consider the subgroup given by \(H=N_ G(ZJ(S))\) where \(S\) is a Sylow \(p\)-subgroup of \(G\) and \(J\) denotes the Thompson subgroup. Under the hypothesis that the elements of order \(p\) in \(G\) have \(p\)-nilpotent centralizers it is shown that the module categories of \(B_ 0(kG)\) and \(B_ 0(kH)\) are equivalent modulo projectives and that the group \(H\) is of such a tight structure that for \(H\) the conclusion can be proved directly. Finally the authors deal with the connections of their results to the theory of module varieties. They prove that every non-projective module in \(B_ 0(kG)\) has non-trivial cohomology if and only if every non-projective periodic module in \(B_ 0(kG)\) has non-trivial cohomology. Also, they state a conjecture in terms of module varieties that would imply the above conjecture for all \(p\).
    0 references
    Tate cohomology groups
    0 references
    modular representations
    0 references
    cohomology theory of finite groups
    0 references
    projective modules
    0 references
    principal block
    0 references
    non-trivial cohomology
    0 references
    sources
    0 references
    vertices
    0 references
    Sylow subgroups
    0 references
    Thompson subgroup
    0 references
    module categories
    0 references
    module varieties
    0 references
    0 references

    Identifiers