Picard groups and class groups of monoid schemes (Q404188)

The authors develop the basic theory of the Picard group of a monoid scheme and of the class group of a normal monoid scheme. Monoid schemes, also known as schemes over the field \({\mathbb F}_1\) with one element, were introduced by \textit{K. Kato} [Am. J. Math. 116, No. 5, 1073--1099 (1994; Zbl 0832.14002)], \textit{A. Deitmar} [Prog. Math. 239, 87--100 (2005; Zbl 1098.14003)] and \textit{A. Connes} and \textit{C. Consani} [J. Algebr. Geom. 20, No. 3, 525--557 (2011; Zbl 1227.14006)] for several reasons. In particular, one can associate a monoid scheme to the fan of every toric variety. The reader willing to learn some basic definitions and facts can consult, for example, the papers of \textit{C. Chu} et al. [Adv. Math. 229, No. 4, 2239--2286 (2012; Zbl 1288.19004)] and of \textit{G. Cortiñas} et al. [J. Reine Angew. Math. 698, 1--54 (2015; Zbl 1331.14050)]. In the first section, the authors study associated primes and primary decompositions in Noetherian monoids. A \textit{monoid} is a (multiplicatively written) commutative semi-group \(A\) with an absorbing element and an identity element, i.e., with elements \(0\) and \(1\) such that \(0 \cdot a = 0\) and \(1 \cdot a = a\), \(\forall \, a \in A\). A non-empty subset \(I\) of \(A\) is an \textit{ideal} if \(AI \subseteq I\). In this case, \(A/I\) denotes the set \((A \setminus I) \cup \{0\}\) endowed with the monoid structure making the surjection \(\pi : A \rightarrow A/I\) collapsing \(I\) to \(\{0\}\) a morphism of monoids. A proper ideal \(\mathfrak{p} \neq A\) of \(A\) is a \textit{prime ideal} if \(S := A \setminus \mathfrak{p}\) is a multiplicatively closed subset of \(A\). The monoid of fractions \(S^{-1}A\) is denoted, in this case, by \(A_{\mathfrak{p}}\). The rest of the definitions are as in commutative ring theory and the authors prove, actually, the analogues of many results valid for commutative rings. In Section 2, the authors study discrete valuation monoids and normal monoids. Let us reproduce some definitions. A monoid \(A\) is \textit{cancellative} if \(ab = ac\) and \(a \neq 0\) implies \(b = c\). In this case, \(\{0\}\) is a prime ideal of \(A\) and the localization \(A_{\{0\}}\) is the \textit{group completion} of \(A\). A cancellative monoid is \textit{normal} if \(\forall \, \alpha \in A_{\{0\}}\), \(\forall \, n \geq 1\), \(\alpha^n \in A\) implies \(\alpha \in A\). The authors show, for example, that every \(1\)-dimensional, Noetherian normal monoid is a discrete valuation monoid, and that a Noetherian normal monoid \(A\) is the intersection of \(A_{\mathfrak{p}}\) as \(\mathfrak{p}\) runs over all height \(1\) primes of \(A\). Most of the facts proven in the first two sections can be also found in the Diploma thesis of U. Kobsa (Regensburg University, 1996) and in [\textit{F. Halter-Koch}, Ideal systems. An introduction to multiplicative ideal theory. Pure and Applied Mathematics, Marcel Dekker. 211. New York, NY: Marcel Dekker (1998; Zbl 0953.13001)]. In the third section, the authors construct the \textit{normalization} of a \textit{partially cancellative} monoid, i.e., of a monoid of the form \(C/I\) where \(C\) is a cancellative monoid. Unless the monoid is cancellative, this normalization turns out to be not a monoid but a \textit{monoid scheme}. Let us recall, for later reference, the definition of such an object and some basic constructions. A \textit{monoid space} is a pair \((X, {\mathcal A})\), where \(X\) is a topological space and \(\mathcal A\) is a sheaf of monoids. If \(A\) is a monoid, \(\text{MSpec}(A)\) denotes the monoid space with \(X\) the set of prime ideals of \(A\) with the Zariski topology and such that the stalk of \(\mathcal A\) at a prime ideal \(\mathfrak{p} \in X\) is \(A_{\mathfrak{p}}\). A monoid scheme is a monoid space looking locally as an \(\text{MSpec}(A)\). If \(A\) and \(B\) are monoids then \(A \vee B := A\times \{0\} \cup \{0\} \times B\) is an ideal of the monoid \(A \times B\) (with componentwise multiplication). The \textit{smash product} \(A \wedge B\) is, by definition, \(A \times B/A \vee B\). It turns out that \(\text{MSpec}(A \wedge B)\) is the product of \(\text{MSpec}(A)\) and \(\text{MSpec}(B)\) in the category of monoid schemes (its underlying set is, actually, \(\text{MSpec}(A) \times \text{MSpec}(B)\)). If \(A = A_0 \vee A_1 \vee \ldots \vee A_n \vee \ldots\) is an \(\mathbb N\)-graded monoid (this means that \(A_i \cap A_j = \{0\}\) for \(i \neq j\) and that \(A_i \cdot A_j \subseteq A_{i+j}\), \(\forall \, i, j\)) then the monoid scheme \(\text{MProj}(A)\) is defined as follows: if \(\mathfrak{p} \in \text{MSpec}(A_0)\) then \(\mathfrak{p} \vee A_1 \vee \ldots \vee A_n \vee \ldots \in \text{MSpec}(A)\), hence \(\text{MSpec}(A_0)\) can be embedded into \(\text{MSpec}(A)\). Then \(\text{MProj}(A) = (X,\mathcal A)\), where \(X = \text{MSpec}(A) \setminus \text{MSpec}(A_0)\) and the stalk of \(\mathcal A\) at a point \(\mathfrak{p} \in X\) is the degree \(0\) part of \(A_{\mathfrak{p}}\). In Section 4, the authors define \textit{Weil divisors} on a normal monoid scheme \(X\). They define \(\text{Div}(X)\) as the free abelian group generated by height \(1\) points of \(X\) and \(\text{Cl}(X)\) as the quotient of \(\text{Div}(X)\) by the subgroup of \textit{principal divisors}. They show, for example, that \(\text{Cl}(X \times Y) = \text{Cl}(X) \times \text{Cl}(Y)\). In the fifth section, the authors study the \textit{Picard group} \(\text{Pic}(X)\) of isomorphism classes of invertible sheaves on a monoid scheme \(X\). Let us recall the definition of a sheaf of \(\mathcal A\)-sets on \((X, \mathcal A)\). If \(A\) is a monoid then an \(A\)-\textit{set} is a pointed set \(M\), with base point \(*\), together with a map \(\theta : A \times M \rightarrow M\) such that \[ (ab)\cdot m = a\cdot (b\cdot m)\, ,\;1 \cdot m = m\, ,\;0 \cdot m = *\, ,\;a \cdot * = *\, . \] If \(M\) and \(N\) are \(A\)-sets then \(M \wedge_A N\) is the quotient of \(M \times N\) by the equivalence relation generated by \((a \cdot m, n) \sim (m, a \cdot n)\), \(a \in A\), \(m \in M\), \(n \in N\). This equivalence relation is compatible with the diagonal action of \(A\) on \(M \times N\), hence \(M \wedge_A N\) has a natural structure of \(A\)-set. It is an analogue of the tensor product. Now, an invertible sheaf on \((X,\mathcal A)\) is a sheaf \(\mathcal L\) of \(\mathcal A\)-sets locally isomorphic to \(\mathcal A\). The set \(\text{Pic}(X)\) of isomorphism classes of invertible sheaves on \(X\) endowed with the operation defined by the tensor product is a group. The authors show, for example, that \(\text{Pic}(\text{MSpec}(A)) = 0\) and that \(\text{Pic}(X \times {\mathbb A}^1) \simeq \text{Pic}(X)\). Here \({\mathbb A}^1\) is \(\text{MSpec}\) of the monoid \(\{T^n\, | \, n \in {\mathbb N}\} \cup \{0\}\). In Section 6, the authors show that if \(X\) is a separated connected locally factorial monoid scheme then every Weil divisor on \(X\) is a Cartier divisor and \(\text{Pic}(X) = \text{Cl}(X)\). They also show that \(\text{Pic}({\mathbb P}^n) = {\mathbb Z}\), where \({\mathbb P}^n\) is \(\text{MProj}\) of the graded monoid \(\{ T_0^{a_0}\ldots T_n^{a_n} \, | \, a_0, \ldots , a_n \in {\mathbb N}\} \cup \{0\}\). In the final section, the authors prove some results and give some counterexamples concerning the Picard group of partially cancellative monoid schemes.