I've recently been reading about primary decomposition, and was browsing the questions here. From this, I know that it is not true that every primary ideal is the power of a prime ideal.

I'm curious about a slight variation inspired by the comments.

Consider $K[X,Y]$, for $K$ a field. For any prime $\mathfrak{p}\subset K[X,Y]$, and any $n\geq 1$, is the ideal $\mathfrak{p}^n$ primary, so that if $ab\in\mathfrak{p}^n$, then either $a\in\mathfrak{p}^n$ or $b^j\in\mathfrak{p}^n$ for some $j$?

I suspect it is indeed true, and it is obviously true when $n=1$. An inductive approach doesn't feel right, so is there another way to reach the conclusion?


1 Answer 1


The answer is yes. Recall the following important fact:

Theorem. (Reid, Undergraduate Commutative Algebra, Proposition page 22) The prime ideals of $k[x,y]$ are as follows:

  1. $0$;
  2. $(f)$, for irreducible $f \in k[x,y]$;
  3. maximal ideals $\mathfrak{m}$.

Now you can easily conclude with the following two facts: (a) a power of a maximal ideal is primary (Atiyah, Macdonald 4.2); (b) in a UFD the power of a principal prime ideal is primary.

I don't know what happens in general in $k[x_1, \dots, x_n]$, when $n \geq 3$. It can be proved that if $\mathfrak{p}$ is a monomial prime in $k[x_1, \dots, x_n]$ then $\mathfrak{p}^m$ is $\mathfrak{p}$-primary for all $m$.

  • $\begingroup$ I doubt this holds for $n\geq 3$. $\endgroup$ Feb 24, 2012 at 20:27
  • $\begingroup$ Thanks Andrea. Do you have a proof or reference to a proof of the second fact (b) stated above? $\endgroup$
    – Kally
    Feb 25, 2012 at 1:16
  • $\begingroup$ It is quite easy. Try to prove it on your own. $\endgroup$
    – Andrea
    Feb 25, 2012 at 13:46
  • $\begingroup$ @Alex: you are right to doubt! See a non-primary square of a prime in Atiyah-Macdonald, page 51, Example 3). $\endgroup$ Sep 3, 2015 at 6:42
  • $\begingroup$ @GeorgesElencwajg That doubt refers to the polynomial ring over a field in (at least) three variables. There is such an example in Northcott, Ideal Theory, page 29, Example 3. $\endgroup$
    – user26857
    Oct 5, 2016 at 21:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .