\(K\)-theory classification of graded ultramatricial algebras with involution (Q2418035)

The heart of the main result of the paper under review can be formulated as ``if \(\Gamma\) is an abelian group, then the graded Grothendieck group \(K^{\mathrm{gr}}_0\) completely classifies graded ultramatricial \(*\)-algebras over a \(\Gamma\)-graded \(*\)-field \(A\) such that (1) each nontrivial graded component of \(A\) has a unitary element (i.e., an element \(a\) such that \(aa^* = a^*a = 1\)) and (2) the zero-component \(A_0\) is \(2\)-proper (\(aa^* + bb^* = 0\) implies \(a = b = 0\) for any \(a, b \in A_0\)) and \(*\)-Pythagorean (for any \(a, b \in A_0\) one has \(aa^* + bb^* = cc^*\) for some \(c \in A_0\))''. Namely, if \(R\) and \(S\) are graded ultramatricial \(*\)-algebras over \(A\) and \(f : K^{\mathrm{gr}}_0 (R) \to K^{\mathrm{gr}}_0 (S)\) is a contractive \(\mathbb{Z}[\Gamma]\)-module isomorphism (i.e., \(f\) preserves the pre-order on \(K^{\mathrm{gr}}_0 (R)\) which has the monoid \(\mathcal{V}^{\mathrm{gr}}(R)\) of graded isomorphism classes of finitely generated graded projective \(R\)-modules as positive cone and \(f\) satisfies \(f([R]) \leq [S]\)), then the authors explicitly construct a graded \(A\)-algebra \(*\)-isomorphism \(\varphi : R \to S\) such that \(K^{\mathrm{gr}}_0 (\varphi) = f\). A \(\Gamma\)-graded ring \(A\) together with an involution \(*: A \to A\) is said to be a \(\Gamma\)-graded \(*\)-ring if \(A_\gamma^* = A_{- \gamma}\) for every \(\gamma \in \Gamma\), where \(A_\gamma^* = \{a^* \mid a^* \in A_\gamma\}\). For the sake of the interested reader, a \(\Gamma\)-graded \(*\)-ring can be equivalently described as a not necessarily unital comodule algebra over the Hopf algebra \(\mathbb{Z}[\Gamma]\), together with an isomorphism of comodule algebras \(*:A \to A^{\mathrm{op}}\) (where the coaction on \(A^{\mathrm{op}}\) is provided by \(a \mapsto a \otimes S(\gamma)\) for all \(a \in A_\gamma\) and for all \(\gamma\in\Gamma\)). A \(\Gamma\)-graded \(*\)-ring \(A\) is a \(\Gamma\)-graded \(*\)-field if it is commutative and every homogeneous element has a multiplicative inverse. A ring \(R\) is a graded ultramatricial \(*\)-algebra over a \(\Gamma\)-graded \(*\)-field \(A\) if \(R\) is a directed union of an ascending sequence of graded matricial \(*\)-subalgebras \(R_1 \subseteq R_2 \subseteq \ldots\) (i.e., graded \(A\)-algebras of the form \(\mathsf{Mat}_{n_1} (A) \oplus \cdots \oplus \mathsf{Mat}_{n_k} (A)\), with a suitable grading) such that the inclusion \(R_n \subseteq R_{n+1}\) is a graded \(*\)-homomorphism (not necessarily unital) for any \(n\). The graded Grothendieck group \(K^{\mathrm{gr}}_0\) is defined as the group completion of the monoid \(\mathcal{V}^{\mathrm{gr}}(R)\) (see [\textit{D. Quillen}, in: Cohomology of groups and algebraic \(K\)-theory. Selected papers of the international summer school on cohomology of groups and algebraic \(K\)-theory, Hangzhou, China, July 1--3, 2007. Somerville, MA: International Press; Beijing: Higher Education Press. 413--478 (2010; Zbl 1198.19001)] for the general construction). It has both a natural \(\mathbb{Z}_2\)-action (which becomes trivial for graded ultramatricial \(*\)-algebras over a graded \(*\)-field) and a natural \(\Gamma\)-action. One says that \(K^{\mathrm{gr}}_0\) completely classifies \(*\)-algebras in a class \(\mathcal{C}\) of graded \(*\)-algebras if any graded \(*\)-algebras \(R\) and \(S\) from \(\mathcal{C}\) are isomorphic as \(*\)-algebras if and only if there is a contractive \((\mathbb{Z}[\Gamma],\mathbb{Z}[\mathbb{Z}_2])\)-bimodule isomorphism \(K^{\mathrm{gr}}_0 (R) \cong K^{\mathrm{gr}}_0 (S)\). We point out that the hypotheses on \(A\) in the main result are strictly connected to the isomorphism being a \(*\)-isomorphism. By considering the group \(\Gamma\) to be trivial and/or by disregarding the involutive structure \(*\), the authors recover a number of know results over more general fields (see Corollaries from 4.7 to 4.11). As a final application, the authors show that if \(E\) and \(F\) are countable, row-finite, no-exit graphs in which every infinite path ends in a sink or a cycle and if \(K\) is a \(2\)-proper and \(*\)-Pythagorean field, then the Leavitt path algebras \(L_K(E)\) and \(L_K(F)\) are isomorphic as graded rings if any only if they are isomorphic as graded \(*\)-algebras (which is an instance of the so-called generalized strong isomorphism conjecture).