4
$\begingroup$

Let $X$ be a $3$-fold, and $f:Y\rightarrow X$ a birational $\mathbb{Q}$-factorial divisorial terminal contraction (of relative Picard number one) contracting a divisor $E\subset Y$ to a point $p\in X$. Is the exceptional divisor $E$ necessarily irreducible?

$\endgroup$
5
$\begingroup$

Since $f:Y\rightarrow X$ is terminal we may write $$K_Y = f^{*}K_X + aE$$ with $a>0$. Let $C\subset Y$ be a curve in the ray contracted by $f$. Then $$K_Y\cdot C = C\cdot f^{*}K_X+aC\cdot E = K_X\cdot f_{*}C+aC\cdot E = aC\cdot E$$ Now, $K_Y\cdot C<0$ yields $C\cdot E < 0$.

Assume that $E = E_1\cup E_2$ has two components and that $Y$ is $\mathbb{Q}$-factorial. Then $E_1, E_2$ are $\mathbb{Q}$-Cartier, $C\cdot E = C\cdot E_1 + C\cdot E_2$ implies that $C\cdot E_1<0$ or $C\cdot E_2<0$.

If $C\cdot E_1<0$ then $C\subset E_1$ and the curves numerically equivalent to $C$ can not cover $E_2$ which therefore is not contracted by $f$.

Similarly, if $C\cdot E_2<0$ then $C\subset E_2$ and the curves numerically equivalent to $C$ can not cover $E_1$ which therefore is not contracted by $f$.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.