The Stacks project

\begin{equation*} \DeclareMathOperator\Coim{Coim} \DeclareMathOperator\Coker{Coker} \DeclareMathOperator\Ext{Ext} \DeclareMathOperator\Hom{Hom} \DeclareMathOperator\Im{Im} \DeclareMathOperator\Ker{Ker} \DeclareMathOperator\Mor{Mor} \DeclareMathOperator\Ob{Ob} \DeclareMathOperator\Sh{Sh} \DeclareMathOperator\SheafExt{\mathcal{E}\mathit{xt}} \DeclareMathOperator\SheafHom{\mathcal{H}\mathit{om}} \DeclareMathOperator\Spec{Spec} \newcommand\colim{\mathop{\mathrm{colim}}\nolimits} \newcommand\lim{\mathop{\mathrm{lim}}\nolimits} \newcommand\Qcoh{\mathit{Qcoh}} \newcommand\Sch{\mathit{Sch}} \newcommand\QCohstack{\mathcal{QC}\!\mathit{oh}} \newcommand\Cohstack{\mathcal{C}\!\mathit{oh}} \newcommand\Spacesstack{\mathcal{S}\!\mathit{paces}} \newcommand\Quotfunctor{\mathrm{Quot}} \newcommand\Hilbfunctor{\mathrm{Hilb}} \newcommand\Curvesstack{\mathcal{C}\!\mathit{urves}} \newcommand\Polarizedstack{\mathcal{P}\!\mathit{olarized}} \newcommand\Complexesstack{\mathcal{C}\!\mathit{omplexes}} \newcommand\Pic{\mathop{\mathrm{Pic}}\nolimits} \newcommand\Picardstack{\mathcal{P}\!\mathit{ic}} \newcommand\Picardfunctor{\mathrm{Pic}} \newcommand\Deformationcategory{\mathcal{D}\!\mathit{ef}} \end{equation*}

10.157 Ascending properties

In this section we start proving some algebraic facts concerning the “ascent” of properties of rings. To do this for depth of rings one uses the following result on ascending depth of modules, see [IV, Proposition 6.3.1, EGA].

reference

Lemma 10.157.1. We have

\[ \text{depth}(M \otimes _ R N) = \text{depth}(M) + \text{depth}(N/\mathfrak m_ RN) \]

where $R \to S$ is a local homomorphism of local Noetherian rings, $M$ is a finite $R$-module, and $N$ is a finite $S$-module flat over $R$.

Proof. In the statement and in the proof below, we take the depth of $M$ as an $R$-module, the depth of $M \otimes _ R N$ as an $S$-module, and the depth of $N/\mathfrak m_ RN$ as an $S/\mathfrak m_ RS$-module. Denote $n$ the right hand side. First assume that $n$ is zero. Then both $\text{depth}(M) = 0$ and $\text{depth}(N/\mathfrak m_ RN) = 0$. This means there is a $z \in M$ whose annihilator is $\mathfrak m_ R$ and a $\overline{y} \in N/\mathfrak m_ RN$ whose annihilator is $\mathfrak m_ S/\mathfrak m_ RS$. Let $y \in N$ be a lift of $\overline{y}$. Since $N$ is flat over $R$ the map $z : R/\mathfrak m_ R \to M$ produces an injective map $N/\mathfrak m_ RN \to M \otimes _ R N$. Hence the annihilator of $z \otimes y$ is $\mathfrak m_ S$. Thus $\text{depth}(M \otimes _ R N) = 0$ as well.

Assume $n > 0$. If $\text{depth}(N/\mathfrak m_ RN) > 0$, then we may choose $f \in \mathfrak m_ S$ mapping to $\overline{f} \in S/\mathfrak m_ RS$ which is a nonzerodivisor on $N/\mathfrak m_ RN$. Then $\text{depth}(N/\mathfrak m_ RN) = \text{depth}(N/(f, \mathfrak m_ R)N) + 1$ by Lemma 10.71.7. According to Lemma 10.98.1 the element $f \in S$ is a nonzerodivisor on $N$ and $N/fN$ is flat over $R$. Hence by induction on $n$ we have

\[ \text{depth}(M \otimes _ R N/fN) = \text{depth}(M) + \text{depth}(N/(f, \mathfrak m_ R)N). \]

Because $N/fN$ is flat over $R$ the sequence

\[ 0 \to M \otimes _ R N \to M \otimes _ R N \to M \otimes _ R N/fN \to 0 \]

is exact where the first map is multiplication by $f$ (Lemma 10.38.12). Hence by Lemma 10.71.7 we find that $\text{depth}(M \otimes _ R N) = \text{depth}(M \otimes _ R N/fN) + 1$ and we conclude that equality holds in the formula of the lemma.

If $n > 0$, but $\text{depth}(N/\mathfrak m_ RN) = 0$, then we can choose $f \in \mathfrak m_ R$ which is a nonzerodivisor on $M$. As $N$ is flat over $R$ it is also the case that $f$ is a nonzerodivisor on $M \otimes _ R N$. By induction on $n$ again we have

\[ \text{depth}(M/fM \otimes _ R N) = \text{depth}(M/fM) + \text{depth}(N/\mathfrak m_ RN). \]

In this case $\text{depth}(M \otimes _ R N) = \text{depth}(M/fM \otimes _ R N) + 1$ and $\text{depth}(M) = \text{depth}(M/fM) + 1$ by Lemma 10.71.7 and we conclude that equality holds in the formula of the lemma. $\square$

Lemma 10.157.2. Suppose that $R \to S$ is a flat and local ring homomorphism of Noetherian local rings. Then

\[ \text{depth}(S) = \text{depth}(R) + \text{depth}(S/\mathfrak m_ RS). \]

Proof. This is a special case of Lemma 10.157.1. $\square$

Lemma 10.157.3. Let $R \to S$ be a flat local homomorphism of local Noetherian rings. Then the following are equivalent

  1. $S$ is Cohen-Macaulay, and

  2. $R$ and $S/\mathfrak m_ RS$ are Cohen-Macaulay.

Proof. Follows from the definitions and Lemmas 10.157.2 and 10.111.7. $\square$

Lemma 10.157.4. Let $\varphi : R \to S$ be a ring map. Assume

  1. $R$ is Noetherian,

  2. $S$ is Noetherian,

  3. $\varphi $ is flat,

  4. the fibre rings $S \otimes _ R \kappa (\mathfrak p)$ are $(S_ k)$, and

  5. $R$ has property $(S_ k)$.

Then $S$ has property $(S_ k)$.

Proof. Let $\mathfrak q$ be a prime of $S$ lying over a prime $\mathfrak p$ of $R$. By Lemma 10.157.2 we have

\[ \text{depth}(S_{\mathfrak q}) = \text{depth}(S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}) + \text{depth}(R_{\mathfrak p}). \]

On the other hand, we have

\[ \dim (R_{\mathfrak p}) + \dim (S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}) \geq \dim (S_{\mathfrak q}) \]

by Lemma 10.111.6. (Actually equality holds, by Lemma 10.111.7 but strictly speaking we do not need this.) Finally, as the fibre rings of the map are assumed $(S_ k)$ we see that $\text{depth}(S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}) \geq \min (k, \dim (S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}))$. Thus the lemma follows by the following string of inequalities

\begin{eqnarray*} \text{depth}(S_{\mathfrak q}) & = & \text{depth}(S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}) + \text{depth}(R_{\mathfrak p}) \\ & \geq & \min (k, \dim (S_{\mathfrak q}/\mathfrak pS_{\mathfrak q})) + \min (k, \dim (R_{\mathfrak p})) \\ & = & \min (2k, \dim (S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}) + k, k + \dim (R_\mathfrak p), \dim (S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}) + \dim (R_{\mathfrak p})) \\ & \geq & \min (k, \dim (S_{\mathfrak q})) \end{eqnarray*}

as desired. $\square$

Lemma 10.157.5. Let $\varphi : R \to S$ be a ring map. Assume

  1. $R$ is Noetherian,

  2. $S$ is Noetherian

  3. $\varphi $ is flat,

  4. the fibre rings $S \otimes _ R \kappa (\mathfrak p)$ have property $(R_ k)$, and

  5. $R$ has property $(R_ k)$.

Then $S$ has property $(R_ k)$.

Proof. Let $\mathfrak q$ be a prime of $S$ lying over a prime $\mathfrak p$ of $R$. Assume that $\dim (S_{\mathfrak q}) \leq k$. Since $\dim (S_{\mathfrak q}) = \dim (R_{\mathfrak p}) + \dim (S_{\mathfrak q}/\mathfrak pS_{\mathfrak q})$ by Lemma 10.111.7 we see that $\dim (R_{\mathfrak p}) \leq k$ and $\dim (S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}) \leq k$. Hence $R_{\mathfrak p}$ and $S_{\mathfrak q}/\mathfrak pS_{\mathfrak q}$ are regular by assumption. It follows that $S_{\mathfrak q}$ is regular by Lemma 10.111.8. $\square$

Lemma 10.157.6. Let $\varphi : R \to S$ be a ring map. Assume

  1. $R$ is Noetherian,

  2. $S$ is Noetherian

  3. $\varphi $ is flat,

  4. the fibre rings $S \otimes _ R \kappa (\mathfrak p)$ are reduced,

  5. $R$ is reduced.

Then $S$ is reduced.

Proof. For Noetherian rings reduced is the same as having properties $(S_1)$ and $(R_0)$, see Lemma 10.151.3. Thus we know $R$ and the fibre rings have these properties. Hence we may apply Lemmas 10.157.4 and 10.157.5 and we see that $S$ is $(S_1)$ and $(R_0)$, in other words reduced by Lemma 10.151.3 again. $\square$

Lemma 10.157.7. Let $\varphi : R \to S$ be a ring map. Assume

  1. $\varphi $ is smooth,

  2. $R$ is reduced.

Then $S$ is reduced.

Proof. Observe that $R \to S$ is flat with regular fibres (see the list of results on smooth ring maps in Section 10.140). In particular, the fibres are reduced. Thus if $R$ is Noetherian, then $S$ is Noetherian and we get the result from Lemma 10.157.6.

In the general case we may find a finitely generated $\mathbf{Z}$-subalgebra $R_0 \subset R$ and a smooth ring map $R_0 \to S_0$ such that $S \cong R \otimes _{R_0} S_0$, see remark (10) in Section 10.140. Now, if $x \in S$ is an element with $x^2 = 0$, then we can enlarge $R_0$ and assume that $x$ comes from an element $x_0 \in S_0$. After enlarging $R_0$ once more we may assume that $x_0^2 = 0$ in $S_0$. However, since $R_0 \subset R$ is reduced we see that $S_0$ is reduced and hence $x_0 = 0$ as desired. $\square$

Lemma 10.157.8. Let $\varphi : R \to S$ be a ring map. Assume

  1. $R$ is Noetherian,

  2. $S$ is Noetherian,

  3. $\varphi $ is flat,

  4. the fibre rings $S \otimes _ R \kappa (\mathfrak p)$ are normal, and

  5. $R$ is normal.

Then $S$ is normal.

Proof. For a Noetherian ring being normal is the same as having properties $(S_2)$ and $(R_1)$, see Lemma 10.151.4. Thus we know $R$ and the fibre rings have these properties. Hence we may apply Lemmas 10.157.4 and 10.157.5 and we see that $S$ is $(S_2)$ and $(R_1)$, in other words normal by Lemma 10.151.4 again. $\square$

Lemma 10.157.9. Let $\varphi : R \to S$ be a ring map. Assume

  1. $\varphi $ is smooth,

  2. $R$ is normal.

Then $S$ is normal.

Proof. Observe that $R \to S$ is flat with regular fibres (see the list of results on smooth ring maps in Section 10.140). In particular, the fibres are normal. Thus if $R$ is Noetherian, then $S$ is Noetherian and we get the result from Lemma 10.157.8.

The general case. First note that $R$ is reduced and hence $S$ is reduced by Lemma 10.157.7. Let $\mathfrak q$ be a prime of $S$ and let $\mathfrak p$ be the corresponding prime of $R$. Note that $R_{\mathfrak p}$ is a normal domain. We have to show that $S_{\mathfrak q}$ is a normal domain. To do this we may replace $R$ by $R_{\mathfrak p}$ and $S$ by $S_{\mathfrak p}$. Hence we may assume that $R$ is a normal domain.

Assume $R \to S$ smooth, and $R$ a normal domain. We may find a finitely generated $\mathbf{Z}$-subalgebra $R_0 \subset R$ and a smooth ring map $R_0 \to S_0$ such that $S \cong R \otimes _{R_0} S_0$, see remark (10) in Section 10.140. As $R_0$ is a Nagata domain (see Proposition 10.156.16) we see that its integral closure $R_0'$ is finite over $R_0$. Moreover, as $R$ is a normal domain it is clear that $R_0' \subset R$. Hence we may replace $R_0$ by $R_0'$ and $S_0$ by $R_0' \otimes _{R_0} S_0$ and assume that $R_0$ is a normal Noetherian domain. By the first paragraph of the proof we conclude that $S_0$ is a normal ring (it need not be a domain of course). In this way we see that $R = \bigcup R_\lambda $ is the union of normal Noetherian domains and correspondingly $S = \mathop{\mathrm{colim}}\nolimits R_\lambda \otimes _{R_0} S_0$ is the colimit of normal rings. This implies that $S$ is a normal ring. Some details omitted. $\square$

slogan

Lemma 10.157.10. Let $\varphi : R \to S$ be a ring map. Assume

  1. $\varphi $ is smooth,

  2. $R$ is a regular ring.

Then $S$ is regular.

Proof. This follows from Lemma 10.157.5 applied for all $(R_ k)$ using Lemma 10.138.3 to see that the hypotheses are satisfied. $\square$


Comments (0)


Post a comment

Your email address will not be published. Required fields are marked.

In your comment you can use Markdown and LaTeX style mathematics (enclose it like $\pi$). A preview option is available if you wish to see how it works out (just click on the eye in the toolbar).

Unfortunately JavaScript is disabled in your browser, so the comment preview function will not work.

All contributions are licensed under the GNU Free Documentation License.




In order to prevent bots from posting comments, we would like you to prove that you are human. You can do this by filling in the name of the current tag in the following input field. As a reminder, this is tag 0336. Beware of the difference between the letter 'O' and the digit '0'.