Nicely semiramified division algebras over Henselian fields. (Q2570965)

Let \((K,v)\) be a Henselian valued field with residue field \(\widehat K\) and value group \(v(K)\). Assume also that \(D\) is a central division \(K\)-algebra of finite dimension \([D:K]\). It is well-known that \(v\) is uniquely, up to equivalence, extendable to a valuation \(v_D\) of \(D\). This also applies to the following results: (i) the value group \(v(D)\) of \((D,v_D)\) is Abelian and contains \(v(K)\) as a subgroup of finite index \(e(D/K)\); (ii) the residue division ring \(\widehat D\) of \((D,v_D)\) is an algebra over \(\widehat K\) of finite dimension \([\widehat D:\widehat K]\); (iii) \([D:K]\) is divisible by \([\widehat D:\widehat K]e(D/K)\). The division \(K\)-algebra \(D\) is called defectless, if \([D:K]=[\widehat D:\widehat K]e(D/K)\); we say that \(D\) is nicely semiramified, if it possesses maximal subfields \(L_1\) and \(L_2\), such that \(L_1\) is totally ramified of radical type and \(L_2\) is inertial over \(K\) [see \textit{B. Jacob, A. Wadsworth}, J. Algebra 128, No. 1, 126-179 (1990; Zbl 0692.16011)]. The paper under review shows that if \(D\) is defectless, then it is nicely semiramified if and only if it is split by inertial and by totally ramified field extensions of \(K\).