Constructing Lie algebra homology classes (Q2572069)

The toral rank conjecture (TRC) for a nilpotent real or complex Lie algebra \({\mathfrak g}\) (always supposed to be finite dimensional) is the assertion that \[ \sum_{i=0}^{n}b_i({\mathfrak g})\geq 2^{\dim Z({\mathfrak g})}, \] where \(b_i({\mathfrak g})\) is the \(i\)th Betti number, i.e. the dimension of \(H^i({\mathfrak g})\), and \(Z({\mathfrak g})=:{\mathfrak z}\) is the center. This conjecture has the following geometric background: calling the toral rank \(\text{rk}\,(M)\) of a manifold \(M\) the dimension of the largest torus which acts freely on \(M\), the toral rank of a nilmanifold \(M=N/\Gamma\) for a nilpotent Lie group \(N\) and a discrete cocompact subgroup \(\Gamma\) is \(\dim Z(N)\), and Steve Halperin conjectured in 1968 that \[ \dim H^*(M)\geq 2^{\text{rk}\,(M)}. \] The TRC has been proven by several people in differents contexts, namely in case \(\dim {\mathfrak z}\leq 5\), \(\dim{\mathfrak g}/ {\mathfrak z}\leq 7\), \(\dim{\mathfrak g}\leq 14\) or in case \({\mathfrak g}\) is \(2\)-step nilpotent. In the article under review, the authors prove that it holds for the class of nilpotent complex Lie algebras \({\mathfrak g}\) which they call metabelian split, i.e. \({\mathfrak g}\) is an abelian extension of an abelian subalgebra \({\mathfrak h}\) by a maximal abelian ideal \({\mathfrak a}\supset {\mathfrak z}\). Indeed, they construct explicitly enough homology classes using the map \[ \Phi_p:\Lambda^p{\mathfrak z}\to\Lambda^{p+m}{\mathfrak g},\quad a_{i_1}\wedge\ldots\wedge a_{i_p}\mapsto a_{i_1}\wedge\ldots\wedge a_{i_p} \wedge h_1\wedge\ldots\wedge h_m, \] where \(a_1,\ldots,a_z\) is some basis of \({\mathfrak z}\), and \(h_1,\ldots,h_m\) is one of \({\mathfrak h}\). In the second part of their work, the authors show that the class of Lie algebras which they study, contains Lie algebras which possess no grading, with arbitrarily big center and nilpotency class. They therefore extend strictly the class of Lie algebras for which the TRC is known to hold.