아무 블로그
derived tensor product over perfect ring 본문
$\newcommand{\Z}{\mathbb{Z}} \newcommand{\Q}{\mathbb{Q}} \newcommand{\R}{\mathbb{R}}
\newcommand{\C}{\mathbb{C}} \newcommand{\Spec}{\mathrm{Spec}\,} \newcommand{\O}{\mathcal{O}} \newcommand{\bb}[1]{\mathbb{#1}} \newcommand{\d}{\mathrm{d}} \newcommand{\F}{\mathbb{F}}
\newcommand{\N}{\mathbb{N}} \newcommand{\P}{\mathbb{P}} \newcommand{\m}{\mathfrak{m}}
\newcommand{\GL}{\mathrm{GL}}
\newcommand{\et}{\mathsf{\acute{e}t}} \newcommand{\lara}[1]{\langle #1 \rangle}
\newcommand{\leris}[1]{\left(#1\right)}
\newcommand{\lerim}[1]{\left\{#1\right\}}
\newcommand{\leril}[1]{\left[#1\right]} \newcommand{\Hom}{\mathrm{Hom}} \newcommand{\Tor}{\mathrm{Tor}}
\newcommand{\Ext}{\mathrm{Ext}}
\newcommand{\holim}{\mathrm{holim}}
\newcommand{\sep}{\mathrm{sep}}
\newcommand{\Gal}{\mathrm{Gal}}
\newcommand{\Mod}{\mathsf{Mod}}
\newcommand{\coker}{\mathrm{coker}}
\newcommand{\Im}{\mathrm{Im}}
\newcommand{\Ker}{\mathrm{Ker}}
\newcommand{\unr}{\mathrm{unr}}
\newcommand{\End}{\mathrm{End}}
\newcommand{\op}{\mathrm{op}}
\newcommand{\Sh}{\mathsf{Sh}}
\newcommand{\Sch}{\mathsf{Sch}}
\newcommand{\Alg}{\mathsf{Alg}}
\newcommand{\p}{\mathfrak{p}}$
여기에 글을 오랜만에 쓰네요. 여길 안 썼던 이유는 단순히 티스토리에서 commutative diagram이 제대로 지원되지 않아서였어요.
Theorem. $A$가 char. p perfect ring이고 $B,C$가 $A$-algebra라고 하자. 그러면 $\Tor^A_i(B,C)^{\frac{1}{p^{\infty}}}=0$이 된다.
재미있는 것 같아요. 이건 perfect ring에겐 derived나 underived나 상관 없음을 말해주는 것 같아요.
Proof) 먼저 filtered colimit argument로 $B,C$가 $A$ 위에서 finite presentation이라고 해도 되고 그럼 $C=A[x_1,\cdots,x_n]/I$라고 해보자고요. 그럼 induction argument를 생각하면 $A[x_1,\cdots,x_{n-1}]$은 $A$ 위에서 flat이니까
$$ \Tor^A_i(B,C)^{\frac{1}{p^{\infty}}}=\Tor^{A[x_1,\cdots,x_n]}(B,C)$$
가 되고, 따라서 $A\to C$는 사실 surjection이라고 해도 되겠지요. 그리고 그 kernel을 $I$라고 하면 $I=(c_1,\cdots,c_n)$이라고 하면 $I'=(c_1,\cdots,c_{n-1})$라고 하면
$$ 0\to A/I'\to A/(c_n)\to A/I=C\to 0$$
에다가 Tor functor를 씌우면 induction hypothesis를 생각하면 $I=(c)$가 principal이라고 해도 될 거예요.
이제 $C/I$에 대한 $i\ge 2$에서의 Tor functor는 펑펑 터지고 따라서 $\Tor^A_1(B,C)^{\frac{1}{p^{\infty}}}=\Tor^A_1(B^{\frac{1}{p^{\infty}}},C^{\frac{1}{p^{\infty}}})=0$임을 증명하면 되는데,
이제 $I^{\frac{1}{p^{\infty}}}\otimes_A B^{\frac{1}{p^{\infty}}}\to B^{\frac{1}{p^{\infty}}}$가 injection이면 모든 게 끝나는데, 이는 $a^{\frac{1}{p^n}}b=0$ for $b\in B^{\frac{1}{p^{\infty}}}$면 $a^{\frac{1}{p^n}-\frac{1}{p^{n+1}}}b=0$임을 증명할 수 있다면
$$ a^{\frac{1}{p^n}}\otimes b=\left(a^{\frac{1}{p^n}-\frac{1}{p^{n+1}}}a^{\frac{1}{p^{n+1}}}\right)\otimes b=a^{\frac{1}{p^{n+1}}}\otimes\left(a^{\frac{1}{p^n}-\frac{1}{p^{n+1}}}b\right)=0$$
이 되어 증명이 완전히 끝나고 남은 건 뒤의 lemma의 argument로 증명하면 되겠지요.
사실, 저것보다 좀 더 강력한 이게 성립해요.
Theorem. 위의 Theorem하고 같은 setting에서 $B,C$가 perfect ring이면 $\Tor^i_A(B,C)=0$ for $i>0$이다.
이걸 증명하는 데는 완전히 같은 argument로 $A\to C$가 surjection이고 그 kernel이 $I^{\frac{1}{p^{\infty}}}=(c^{\frac{1}{p^{\infty}}})$꼴이라고 가정 가능하고 다음 lemma를 보자고요.
Lemma. $I^{\frac{1}{p^{\infty}}}=(c,c^{\frac{1}{p}},c^{\frac{1}{p^2}},\cdots)$는 다음 direct limit
$$ A\overset{c^{1-\frac{1}{p}}}{\longrightarrow}A\overset{c^{\frac{1}{p}-\frac{1}{p^2}}}{\longrightarrow}A\overset{c^{\frac{1}{p^2}-\frac{1}{p^3}}}{\longrightarrow}\cdots$$
로 표현될 수 있다.
Proof) 저거 자체는 직관적으로 저 direct limit가 $\left\{\sum \frac{a_i}{c^{1-\frac{1}{p^i}}}\right\}$꼴로 생겨야 하고, 그럼 $c$를 곱하면 바로 $I$가 되지요.
좀 더 엄밀하게는, $\lim(A\to A\to \cdots)\to I$를 $n$번째 원소를 $a$라고 하면 이걸 $ac^{\frac{1}{p^n}}$으로 보낸다고 생각할께요.(저 위에 원소들은 모두 $0$번째 $A$에 들어가 있다고 가정할 때의 모습이니까, 다시 원래대로 돌아오게 하려면 이게 나와요) 그럼 이건 $n=0$이라고 하면 surjection이고 injection인 건 $ac^{\frac{1}{p^n}}=0$이면 $ac^{\frac{1}{p^n}-\frac{1}{p^m}}=0$ for some $m>n$이 되도록 하면 direct limit의 concrete definition으로 증명이 끝나고 그럼 $ac^{\frac{1}{p^n}}=0$의 양변에 $\frac{1}{p}$제곱을 씌우면 ($A$가 perfect라 가능해요) $a^{\frac{1}{p}}c^{\frac{1}{p^{n+1}}}=0$이 되고 따라서 양변에 $a^{\frac{p-1}{p}}$를 곱하면 $ac^{\frac{1}{p^{n+1}}}=0$이니까 여기에 다시 $c^{\frac{p-2}{p^{n+1}}}$을 양변에 곱하면 되겠네요.
이 lemma가 의미하는 것은 간단히 $I^{\frac{1}{p^{\infty}}}$가 flat $A$-module이란 거예요. 따라서
$$ 0\to I^{\frac{1}{p^{\infty}}}\to A\to C\to 0$$
이란 exact sequence를 생각하면 $\Tor^{i+1}_A(B,C)=\Tor^i_A(B,I^{\frac{1}{p^{\infty}}})$일 테니까 증명이 끝나지요.
이 증명을 직관적으로 생각해내는 것은 생각보다 어렵지 않아요. 먼저 reduction 과정은 아무나 할 수 있고 $cA$는 당연히 $A$ 위에서 flat이고 그럼 저 direct limit는 $c^{\frac{1}{p^n}}A$들의 union이라고 생각하면 되겠지요. 다만 $i=1$의 case가 조금 어려워보이긴 해요.
여기 증명은 https://www.math.uni-bonn.de/people/scholze/Witt.pdf에서 퍼왔어요.
'수학 > 대수' 카테고리의 다른 글
Ring without 1 (0) | 2020.03.10 |
---|