Are sufficient conditions for conservativity of minimal quantum semigroups necessary? (Q2435845)

A construction of a minimal pointwise weak\(^*\)-continuous semigroup of operators on \(B(h)\) (where \(h\) is a Hilbert space), developed by \textit{F. Fagnola} and \textit{A. M. Chebotarev} (see for example [J. Funct. Anal. 153, No. 2, 382--404 (1998; Zbl 0914.47040)]), takes as a starting point a certain form generator \(\mathcal{L}\) built of a generator \(G\) of a strongly continuous contraction semigroup on \(h\) and a family of operators \(L_l\) with domains containing the domain of \(G\). If there exists a (potentially unbounded, but sufficiently well behaved in terms of its domain) operator \(C\) such that \(\sum_l L_l^* L_l \leq C\) and \(\mathcal{L}(C)\leq bC\) for some \(b>0\), the resulting minimal semigroup is conservative, i.e.\ preserves the unit, which further implies a certain uniqueness of the construction. In the paper under review the authors construct an example of a form generator leading to a non-conservative minimal semigroup on \(B(h)\), essentially combining a jump process in the `momentum picture' and Brownian motion in the `position picture'. The example shows that the sufficient assumptions mentioned above are indeed necessary to obtain conservativity.