MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

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


I guess

$$H^2_{(x,y)}\left(\frac{\Bbb Z[x,y]}{(5x+4y)}\right)=0$$

We know that if $\operatorname{Supp} H^i_I(M)‎\subseteq V(I)\cap \operatorname{Supp}(M)$, then $$\operatorname{Supp} H^2_{(x,y)}\frac{\Bbb Z[x,y]}{(5x+4y)})\subseteq V((x,y))\cap V((5x+4y))=V((x,y))=‎\lbrace(x,y) \rbrace\cup ‎\lbrace‎‎(x,y,p) \rbrace‎‎$$ where $p$ is prime number.

If we could show that for every $P\in V((x,y))$, $$\left(H^2_{(x,y)}\left(\frac{\Bbb Z[x,y]}{(5x+4y)}\right)\right)_P=0$$

then it is done.

background: $H^i_I(M)$ means $i$-th local cohomology module of $M$ with respect to ideal $I$. $V(I)=\lbrace P\in Spec(R); I\subseteq P\rbrace $ and $\operatorname{Supp}(M)=\lbrace P\in \operatorname{Spec}(R) ; M_p\neq 0\rbrace$and $M_P$ means $M$ localized at prime ideal $P$. Furthermore $\operatorname{Supp}(R/I)=V(I)$.

share|cite|improve this question
Do you mean $M=S/(xu+yv)$ ? Also I don't understand, why your example fits into this pattern. – Ralph Dec 5 '12 at 10:35
Yes.$M=S/(xu+yv)$. – Angel Dec 5 '12 at 10:39
In your example, the vanishing takes place because of a general vanishing result of Grothendieck. See 6.1.2 in "Local cohomology..." by Brodmann and Sharp (Cambdridge Univ. Press). Like Ralph, I don't understand how your example is a special case. – Damian Rössler Dec 5 '12 at 12:39
How my example be solved by 6.1.2? – Angel Dec 5 '12 at 13:47
I am sorry, I miscalculated: 6.1.2 implies only that $H^2_{(x,y)}({\bf Q}[x,y]/(5x+4y))$. – Damian Rössler Dec 5 '12 at 15:44
up vote 2 down vote accepted

$P=(x,y,p)$ implies $P\cap\mathbb Z=p\mathbb Z$ (here $p$ is a prime or $0$). As localization commutes with local cohomology $$H^2_{(x,y)}\left(\frac{\mathbb{Z}[x,y]}{(5x+4y)}\right)_P\simeq H^2_{(x,y)}\left(\frac{\mathbb{Z}[x,y]_P}{(5x+4y)}\right).$$ But $\mathbb Z[x,y]_P\simeq\mathbb Z_{(p)}[x,y]_{\overline{P}}$, where $\overline{P}$ is the extension of $P$ to $\mathbb Z_{(p)}[x,y]$. Now let's see what happens with the involved ideals via this isomorphism: $(5x+4y)$ goes to $(5x+4y)\mathbb Z_{(p)}[x,y]_{\overline{P}}$ and note that $5$ or $4$ (or both) are invertible in $\mathbb Z_{(p)}$, so the factor ring $\mathbb Z_{(p)}[x,y]_{\overline{P}}/(5x+4y)\mathbb Z_{(p)}[x,y]_{\overline{P}}$ is isomorphic to $\mathbb Z_{(p)}[t]_Q$, where $Q\cap \mathbb Z_{(p)}=p\mathbb Z_{(p)}$. Local cohomology is independent of base ring, so finally we arrive to $H^2_{(t)}(\mathbb Z_{(p)}[t]_Q)=0$ (since the local cohomology in a principal ideal is zero from $2$ onwards).

share|cite|improve this answer
Dear YACP: As you can see on the upper margin of this page: "put backticks around any math that contains underscores or asterisks." – Mahdi Majidi-Zolbanin Dec 8 '12 at 18:30

I wasn't able to use the method you suggest. Here is a different proof. Let $M$ be any ${\bf Z}[x,y]$-module. We have (see Brodmann and Sharp, Th. 1.3.8), $$ H^2_{(x,y)}(M)=\varinjlim_n \operatorname{Ext}^2({\bf Z}[x,y]/(x,y)^n,M) $$ where the limit is an inductive limit. Now we have $$ \varinjlim_n \operatorname{Ext}^2({\bf Z}[x,y]/(x,y)^n,M)=\varinjlim_n \operatorname{Ext}^2({\bf Z}[x,y]/(x^n,y^n),M) $$ because for any $n$, we have $(x^n,y^n)\subseteq (x,y)^n$ and $(x,y)^{2n}\subseteq (x^n,y^n)$. Now the ideal $(x^n,y^n)$ defines a regular closed immersion into $\operatorname{Spec} {\bf Z}[x,y]$, so that we have the "fundamental local isomorphism" (see Hartshorne, Residues and duality, Prop. 7.2): ... EDIT: the end of the original argument was wrong. I don't know how to make it work.

share|cite|improve this answer
We should use property of $(5x+4y)$ under localization at $P$. Lichtenbaum,-Hartshorne vanishing Theorem may be useful. – Angel Dec 7 '12 at 14:20
I don't know what are you talking about, but what I said referred to you comment, that is, $H_{(x,y)}^2(\mathbb Z[x,y])=0$ implies, by localizing to $Z−\{0\}$, that the similar local cohomology over $\mathbb Q$ is $0$ and this is false! – user26857 Dec 7 '12 at 19:02
@YACP: I understood Angel's comment as meaning exactly what you are saying; ie Angel explained with his terse vanishing statement that my answer was flawed (which is true), because my answer implied that (incorrect) vanishing. – Damian Rössler Dec 8 '12 at 10:22

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.