MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $\mathbb G_a$ denote the additive group of complex numbers.

Definition: Let $V \subset Y$ be a dense open subset of the affine variety $Y$ and $\pi : P \longrightarrow V$ a $\mathbb G_a$-principal bundle. An affine extension is an affine $\mathbb G_a$-variety $\hat P$ together with a morphism $ \hat \pi: \hat P \longrightarrow Y$ and an equivariant open embedding $ \iota: P \hookrightarrow \hat P$, such that the diagram

$$ \begin{array}{ccc} P & \hookrightarrow & \hat P \\\ \downarrow & & \downarrow \\\ V & \hookrightarrow & Y \end{array} $$ is commutative and $\iota (P)= \hat \pi^{-1}(V).$

I am interested in affine extensions of the trivial $\mathbb G_a$-bundle over the affine plane punctured at the origin, i.e. $\mathbb A^2_*:=Sp(\mathbb C[x,y])\setminus\{\mathbf o\}$ with $\mathbf o:=(x,y)$, and I have the following description:

Proposition: If $P \longrightarrow \mathbb A_*^2$ is the trivial bundle, then any affine extension $\hat P \longrightarrow \mathbb A^2$ is of the form $$ \hat P = Sp (A), $$ where $$ A= \bigoplus_{\nu=0}^\infty \mathfrak m_\nu t^\nu \subset \mathbb C [x,y,t], $$ with a decreasing sequence $(\mathfrak m_\nu)$, $\nu \in \mathbb N$, of ideals $\mathfrak m_\nu \subset \mathbb C [x,y]$, such that

1. $ \mathfrak m_\nu \cdot \mathfrak m_\lambda \subset \mathfrak m_{\nu + \lambda} $ for all $\nu, \lambda\in \mathbb N$,

2. $\mathfrak m_0= \mathbb C [x,y]$, and

3. $V(\mathfrak m_\nu) \subset \mathbf o$ for $\nu > 0$.

On the other hand, every finitely generated $\mathbb C$-algebra of that form defines an affine extension of the trivial bundle.

Question: Does somebody know of a criterion on the sequence $(\mathfrak m_\nu)$$_{\nu \in \mathbb N}$ so that $A$ becomes normal?

Examples: a) If $\mathfrak m_\nu= \mathbb C [x,y]$ for all $\nu$ we have $\hat P \cong \mathbb A^2 \times \mathbb G_a$.

b) If $\mathfrak m_\nu =(x^m,y^n)^\nu$, we have $ A = \mathbb C [x,y,x^mt, y^nt]. $

In the second example, one can see that $A$ is normal for instance if $\mathfrak m_\nu=(x^2,y)^\nu$, but not if $\mathfrak m_\nu=(x^2,y^2)^\nu$ -- in the latter case I think the normalization would be defined by the sequence $\mathfrak m_\nu=(x^2,xy,y^2)^\nu$. Since $A = \mathbb C [x,y,x^mt, y^nt]\cong\mathbb C[x,y,u,v]/(x^mv-y^nu)$ in example b), $Sp(A)$ is a hypersurface in $\mathbb C^4$, so normality is equivalent to singularities being of codimension at least two.

Edit: If the question is difficult in general, I am also interested in the following special case: For which monomial ideals $\mathfrak m\subset\mathbb C[x,y]$, is the ring defined by the sequence $\mathfrak m^\nu$ normal? In this situation I would expect something like: $A$ is normal iff the support of $\mathfrak m^\nu$ consists of all lattice points in the convex hull of the support of $\mathfrak m^\nu$ in $\mathbb R^2$. Here the support of $\mathfrak m^\nu$ is the set of pairs $(k,l)\in\mathbb N^2$ such that $x^ky^l\in\mathfrak m^\nu$.

share|cite|improve this question

Hi Isac, One simple criterion can be given which is analogous to Rees' valuations corresponding to ideals. Set $R := \mathbb{C}[x,y]$, define $\nu: R \to \mathbb{N} \cup \infty$ as $\nu(f) := \max\lbrace k: f \in m_k\rbrace$. Then I believe $A$ is normal if $\nu(f^k) = k\nu(f)$ for all $f \in R$. For other criteria, I would look into Kei-ichi Watanabe's articles.

Edit: Here is a general approach.

Claim 1: $A$ is integrally closed iff every homogeneous (with respect to the grading of $A$) element in $\mathbb{C}[x,y,t]$ (i.e. an element of the form $f(x,y)t^k$) which is integral over $A$ is in $A$.

By Claim 1 the integral equation of $ft^k$ (for $f \in R$) over $A$ be of the form $z^d + \sum_{i=0}^d g_ez^{d-e} = 0$ for some $g_e \in m_{ek}$. This proves Claim 2 below.

Claim 2: $A$ is integrally closed iff $\bar m_k := m_k$ for all $k \geq 1$, where

$\bar m_k := \lbrace f \in m_k: f^d + \sum_{i=0}^d g_ez^{d-e} = 0$ for some $d \geq 0$ and $g_1, \ldots, g_d \in m_{ek}\rbrace$.

In the special case that $m_k$ is a monomial ideal for each $k$, it suffices to prove the integral condition only for monomials in $m_k$. And in the special special case that $m_k = m^k$ for some monomial ideal $m$, it follows that your claim is true, i.e. $A$ is integrally closed iff the support of $m^k$ contains all the monomials in the cone spanned by its monomials.

share|cite|improve this answer
Thanks for the answer! The condition seems to be very strong, as formulated, and I don't really see how it is related to normality. Did you mean the condition should hold for all elements with $\nu(f)\geq 1$? Otherwise, if for instance we take $\mathfrak m$ to be a monomial ideal which contains a power of $x$, but not $x$ itself, and take the ideal sequence to be the powers of $\mathfrak m$, the condition fails (with $f=x$, since $\nu(f)=0$) although $A$ may very well be normal (as in example b)). What I'd really like is of course a condition which is both necessary and sufficient... – Isac Hedén Jan 24 '13 at 14:07
... Did you have a particular article by Kei-ichi Watanabe i mind? Thanks again for your answer! – Isac Hedén Jan 24 '13 at 14:09
Hi Isac, A look at Chapter 5 of might help. – auniket Jan 24 '13 at 17:44

The $\nu$-formulation is a little off, you need to incorporate $t$ as well. Let $ord(\alpha) = \sup\lbrace k: \alpha \in m_k t^k\rbrace$, and let $\overline\nu(f) = \limsup\lbrace ord(f^k)/k : k \in N\rbrace$ (related to Section 6.9 in Huneke-Swanson book). At least if $m_k = I^k$ for some ideal $I$, then for all $f$, if $\overline \nu(f) = ord (f)$, then $A$ is normal.

In a two-dimensional regular local ring or two-dimensional polynomial ring such as here, $k[x,y][It]$ is integrally closed iff $I$ is integrally closed (Chapter 14 in Huneke-Swanson).

share|cite|improve this answer
Thanks for the answer! It looks like a good criterion for the case where $A$ is a Rees algebra. Unfortunately this is not always the case though, since we for instance could modify A by repeating every ideal once, i.e. take the ideal sequence $m_1,m_1,m_2, m_2,m_3,m_3,\ldots$ instead of $m_1,m_2,m_3,\ldots$ where all the $m_i$ are powers of a monomial ideal $m_1$ -- without violating the three conditions that the ideal sequence should satisfy. Could there be a way of associating a Rees algebra to $A$ which has the property of being normal exactly when $A$ is, e.g. by means of a Rees valuation? – Isac Hedén Jan 26 '13 at 18:08

First of all I don't see that $A$ is finitely generated. For this you need equality in 1. for $\lambda,\nu$ big enough.

Secondly, your special case of monomial ideals: Here $A$ is triply graded and therefore the ring of functions on an affine toroidal variety. Hence it is normal iff the monoid $M$ of all $(k,l,\nu)\in\mathbb N^3$ with $x^ky^lt^\nu\in A$ is saturated. This means that if $M_{\mathbb R}$ is the convex cone spanned by $M$ then $M=M_{\mathbb R}\cap{\mathbb Z}^3$.

Conversely you can generate all examples this way. Since $M_{\mathbb R}$ contains $(1,0,0)$ and $(0,0,1)$ it suffices to consider the slice $M_{\mathbb R}\cap\{\nu=1\}$. Let $P$ be its projection into the $(x,y)$-plane. Then $P$ is of the form $$ P=\text{convex hull}(p_1,\ldots,p_s)+{\mathbb R}_{\ge0}^2 $$ with $p_i\in{\mathbb Q}_{\ge0}^2$ and one of the $p_i$ is a multiple of $(1,0)$ and $(0,1)$, respectively. In other words, $P$ looks like a Newton polygon. Another way to describe $P$ is by means of linear inequalities $$ a_ix+b_iy\ge c_i $$ where alle $a_i,b_i,c_i\in{\mathbb Q}_{\ge0}$ and $x\ge0$, $y\ge0$ are among these inequalities.

Now put $$ \mathfrak m_\nu:=\langle x^ky^l\mid(k,l)\in\nu P\rangle $$ and you get lots of normal examples.

I don't think that this criterion is equivalent to each slice $\nu=\text{fixed}$ is saturated.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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