A note on Itoh (\(\operatorname{e}\))-valuation rings of an ideal (Q683500)

Let \(I\) be a regular proper ideal of a Noetherian ring \(R\), \(t\) be an indeterminate over \(R\), \({\mathbf R} = R[u, tI]\) be the Rees ring of \(R\) with respect to \(I\), \((W_1, Q_1), \dots , (W_n, Q_n)\) be the Rees valuation rings of \(uR\) for \(j = 1, \dots , n\), \(uW_j = Q_j^{e_j}\) (so \(e_1, \dots , e_n\) are the Rees integers of \(I\)), and \(e \geq 2\) be an integer. Also, let \({\mathbf S} = {\mathbf R}[u^{\frac{1}{e}}]\), \({\mathbf S}'\) be the integral closure of \({\mathbf S}\) in its total quotient ring, \({\mathbf T} = {\mathbf S}' \cap R[u^{\frac{1}{e}}, t^{\frac{1}{e}}]\), and \({\mathbf r} = u^{\frac{1}{e}}{\mathbf T}\). In [\textit{S. Itoh}, J. Algebra 117, No. 2, 390--401 (1988; Zbl 0653.13003)], Itoh proved the following interesting results (1.1)--(1.3) when \(e\) is an arbitrary common multiple of \(e_1, \dots , e_n\): (1.1) \({\mathbf r}\) is a radical ideal, so the Rees integers of \({\mathbf r}\) and of \((u^{\frac{1}{e}}{\mathbf S})_a\), the integral closure of \(u^{\frac{1}{e}}{\mathbf S}\) in \({\mathbf S}\), are all equal to one. (1.2) There is a one-to-one correspondence between the Rees valuation rings \((V^*, N^*)\) of \({\mathbf r}\) and the Rees valuation rings \((W, Q)\) of \(u{\mathbf R}\); namely, if \(F(u)\) is the quotient field of \(W\), then \(V^*\) is the integral closure of \(W\) in \(F(u^{\frac{1}{e}})\). (1.3) Let \((V^*, N^*)\) and \((W, Q)\) be corresponding Rees valuation rings of \({\mathbf r}\) and \(u{\mathbf R}\), respectively, as in (1.2 ), so \(W = W_j\) for some \(j \in \{1, \dots , n\}\). Then \(QV^* = N^{*\frac{e}{e_j}}\), so the ramification index of \(V ^*\) relative to \(W\) is equal to \(\frac{e}{e_j}\). In fact, the only part of these three results that Itoh specifically stated was that \({\mathbf r}\) is a radical ideal when \(e\) is the least common multiple of \(e_1, \dots , e_n\), while his proof of this result essentially shows that (1.1)-(1.3) hold. In this paper under review, the authors derived some additional nice properties of the Rees valuation rings of ideals like \({\mathbf r}\). They first called the Rees valuation rings of \({\mathbf r} = u^{\frac{1}{e}}{\mathbf T}\) the Itoh (\(e\))-valuation rings of \(I\), that is, the Itoh (\(e\))-valuation rings of \(I\) are the rings \(({\mathbf T}/z)_{(p/z)}\), where \(p\) varies over the (height one) associated prime ideals of \({\mathbf r}\) and \(z\) is the (unique) minimal prime ideal in \({\mathbf T}\) that is contained in \(p\). Then, among other things, they showed: (1) \({\mathbf r}\) is a radical ideal if and only if \(e\) is a common multiple of \(e_1, \dots , e_n\). (2) There is a one-to-one correspondence between the Itoh (\(e\))-valuation rings \((V^*, N^*)\) of \(I\) and the Rees valuation rings \((W, Q)\) of \(uR[u, tI]\); namely, if \(F(u)\) is the quotient field of \(W\), then \(V^*\) is the integral closure of \(W\) in \(F(u^{\frac{1}{e}})\). (3) If \((V ^*, N^*)\) and \((W, Q)\) are corresponding valuation rings, as in (2), then \(V^*\) is a finite integral extension domain of \(W\), and \(W\) and \(V^*\) satisfy the Fundamental Equality with no splitting. Also, if \(uW = Q^k\) for some integer \(k \geq 0\), and if the greatest common divisor of \(k\) and \(e\) is \(d\), and \(c\) is the integer such that \(cd = e\), then \(QV^* = N^{*c}\) and \([(V^*/N^*) : (W/Q)] = d\). Further, if \(uW = Q^k\) and \(e = qk\) is a multiple of \(k\), then there exists a unit \(\theta_k \in V^*\) such that \(V^* = W[\theta_k, u^{\frac{1}{e}}]\) is a finite free integral extension domain of \(W\), \(QV^* = N^{*q}, N^* = u^{\frac{1}{e}}V^*\), and \([V^* : W] = e\).