The Stacks project

Lemma 93.17.6. Let $X$ be a scheme over a field $k$. Assume

  1. $X$ is proper over $k$,

  2. $X$ is a local complete intersection over $k$,

  3. $X$ has dimension $\leq 1$, and

  4. $X \to \mathop{\mathrm{Spec}}(k)$ is smooth except at finitely many points.

Then there exists a flat projective morphism $Y \to \mathop{\mathrm{Spec}}(k[[t]])$ whose generic fibre is smooth and whose special fibre is isomorphic to $X$.

Proof. Observe that $X$ is Cohen-Macaulay, see Algebra, Lemma 10.135.3. Thus $X = X' \amalg X''$ with $\dim (X') = 0$ and $X''$ equidimensional of dimension $1$, see Morphisms, Lemma 29.29.4. Since $X'$ is finite over $k$ (Varieties, Lemma 33.20.2) we can find $Y' \to \mathop{\mathrm{Spec}}(k[[t]])$ with special fibre $X'$ and generic fibre smooth by Lemma 93.17.4. Thus it suffices to prove the lemma for $X''$. After replacing $X$ by $X''$ we have $X$ is Cohen-Macaulay and equidimensional of dimension $1$.

We are going to use deformation theory for the situation $\Lambda = k \to k$. Let $p_1, \ldots , p_ r \in X$ be the closed singular points of $X$, i.e., the points where $X \to \mathop{\mathrm{Spec}}(k)$ isn't smooth. For each $i$ we pick an integer $n_ i$ and a flat, essentially of finite type ring map

\[ k[[t]] \longrightarrow B_ i \]

with $B_ i/tB_ i \cong \mathcal{O}_{X, p_ i}$ such that $t^{n_ i}$ is in the $1$st Fitting ideal of $\Omega _{B_ i/k[[t]]}$. This is possible by Lemma 93.17.5. Observe that the system $(B_ i/t^ nB_ i)$ defines a formal object of $\mathcal{D}\! \mathit{ef}_{\mathcal{O}_{X, p_ i}}$ over $k[[t]]$. By Lemma 93.16.3 the map

\[ \mathcal{D}\! \mathit{ef}_ X \longrightarrow \prod \nolimits _{i = 1, \ldots , r} \mathcal{D}\! \mathit{ef}_{\mathcal{O}_{X, p_ i}} \]

is a smooth map between deformation categories. Hence by Formal Deformation Theory, Lemma 90.8.8 there exists a formal object $(X_ n)$ in $\mathcal{D}\! \mathit{ef}_ X$ mapping to the formal object $\prod _ i (B_ i/t^ n)$ by the arrow above. By More on Morphisms of Spaces, Lemma 76.43.5 there exists a projective scheme $Y$ over $k[[t]]$ and compatible isomorphisms $Y \times _{\mathop{\mathrm{Spec}}(k[[t]])} \mathop{\mathrm{Spec}}(k[t]/(t^ n)) \cong X_ n$. By More on Morphisms, Lemma 37.12.4 we see that $Y \to \mathop{\mathrm{Spec}}(k[[t]])$ is flat. Since $X$ is Cohen-Macaulay and equidimensional of dimension $1$ we may apply Lemma 93.17.1 to check $Y$ has smooth generic fibre1. Choose $n$ strictly larger than the maximum of the integers $n_ i$ found above. It we can show $t^{n - 1}$ is in the first Fitting ideal of $\Omega _{X_ n/S_ n}$ with $S_ n = \mathop{\mathrm{Spec}}(k[t]/(t^ n))$, then the proof is done. To do this it suffices to prove this is true in each of the local rings of $X_ n$ at closed points $p$. However, if $p$ corresponds to a smooth point for $X \to \mathop{\mathrm{Spec}}(k)$, then $\Omega _{X_ n/S_ n, p}$ is free of rank $1$ and the first Fitting ideal is equal to the local ring. If $p = p_ i$ for some $i$, then

\[ \Omega _{X_ n/S_ n, p_ i} = \Omega _{(B_ i/t^ nB_ i)/(k[t]/(t^ n))} = \Omega _{B_ i/k[[t]]}/t^ n\Omega _{B_ i/k[[t]]} \]

Since taking Fitting ideals commutes with base change (with already used this but in this algebraic setting it follows from More on Algebra, Lemma 15.8.4), and since $n - 1 \geq n_ i$ we see that $t^{n - 1}$ is in the Fitting ideal of this module over $B_ i/t^ nB_ i$ as desired. $\square$

[1] Warning: in general it is not true that the local ring of $Y$ at the point $p_ i$ is isomorphic to $B_ i$. We only know that this is true after dividing by $t^ n$ on both sides!

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 0E7Y. Beware of the difference between the letter 'O' and the digit '0'.