Let $(R,m)$ be a Noetherian local domain whose integral closure $S$ is too. Also assume that $S$ is module-finite over $R$.

Let $x\in m^k\setminus m^{k+1}$ and $u\in S^\times$ such that $ux \in R$. Is $ux$ necessarily in $m^k\setminus m^{k+1}$?

The case I care about is when $R$ is an affine semigroup ring over $\Bbb C$ with maximal (multigraded) homogeneous ideal $m$ (or rather the localization of this stuff at $m$).


In general, this can be wrong. Consider $R=\mathbb{R}[[x,y]]/(x^2+y^4)$ (or $\mathbb{R}[x,y]_{(x,y)}/(x^2+y^4)$). Then $u=x/y^2$ is in the integral closure $S$ and it is a unit. $y^2\in\mathfrak{m}^2$, $uy^2=x\in\mathfrak{m}$, but not in its square.

  • $\begingroup$ For posterity: Here's why this answer satisfies the requirements on $S$: The $\Bbb R$-algebra homomorphism $R[u]\to \Bbb C[[y]]$ taking $y$ to itself and $u$ to $i$ is an isomorphism. Since $\Bbb C[[y]]$ is a DVR, it's integrally closed, and therefore $R[u]=S$. We win. $\endgroup$ – Avi Steiner Jun 24 '17 at 18:44
  • $\begingroup$ The reason this counterexample exists is non-standard grading. I think that if $A$ is a finitely generated standard-graded algebra over a field (i.e. $A$ contains a field $k$ consisting of its homogeneous degree zero elements, and $A$ is generated by degree one elements over $k$), and, $\frak m$ its homogeneous maximal ideal, and $R=A_{\frak m}$, then the answer to your question should be 'yes'. In the example above, $A = \mathbb R[x,y] / (x^2 + y^4)$, there is no way to give a grading to $A$ in such a way that it is generated over $\mathbb R$ by degree one elements; deg(x) must be 2deg(y) $\endgroup$ – Neil Epstein Jun 26 '17 at 15:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.