Lemma 10.138.14. Let $R \to S$ be a smooth ring map. Then there exists a subring $R_0 \subset R$ of finite type over $\mathbf{Z}$ and a smooth ring map $R_0 \to S_0$ such that $S \cong R \otimes _{R_0} S_0$.
Proof. We are going to use that smooth is equivalent to finite presentation and formally smooth, see Proposition 10.138.13. Write $S = R[x_1, \ldots , x_ n]/(f_1, \ldots , f_ m)$ and denote $I = (f_1, \ldots , f_ m)$. Choose a right inverse $\sigma : S \to R[x_1, \ldots , x_ n]/I^2$ to the projection to $S$ as in Lemma 10.138.5. Choose $h_ i \in R[x_1, \ldots , x_ n]$ such that $\sigma (x_ i \bmod I) = h_ i \bmod I^2$. Since $x_ i - h_ i \in I$, there exist $b_{ij} \in R[x_1, \ldots , x_ n]$ such that
The fact that $\sigma $ is an $R$-algebra homomorphism $R[x_1, \ldots , x_ n]/I \to R[x_1, \ldots , x_ n]/I^2$ is equivalent to the condition that
for certain $a_{kl} \in R[x_1, \ldots , x_ n]$. Let $R_0 \subset R$ be the subring generated over $\mathbf{Z}$ by all the coefficients of the polynomials $f_ j, h_ i, a_{kl}, b_{ij}$. Set $S_0 = R_0[x_1, \ldots , x_ n]/(f_1, \ldots , f_ m)$, with $I_0 = (f_1, \ldots , f_ m)$. Since the second displayed equation holds in $R_0[x_1, \ldots , x_ n]$ we can let $\sigma _0 : S_0 \to R_0[x_1, \ldots , x_ n]/I_0^2$ be the $R_0$-algebra map defined by the rule $x_ i \mapsto h_ i \bmod I_0^2$. Since the first displayed equation holds in $R_0[x_1, \ldots , x_ n]$ we see that $\sigma _0$ is a right inverse to the projection $R_0[x_1, \ldots , x_ n] / I_0^2 \to R_0[x_1, \ldots , x_ n] / I_0 = S_0$. Thus by Lemma 10.138.5 the ring $S_0$ is formally smooth over $R_0$. $\square$
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Your email address will not be published.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).
All contributions are licensed under the GNU Free Documentation License.
Comments (2)
Comment #9094 by Christian Merten on
Comment #9177 by Stacks project on
There are also: