2
$\begingroup$

Let $A$ and $B$ be $2n$-dimensional complex vector bundles and $\det A=\Lambda^{2n}(A)$ and $\det B=\Lambda^{2n}(B)$.

Can you prove $A\cong B $ if and only if $\det A\cong \det B $? Is it a correct proposition?

Note: Here $\cong$ is a bundle isomorphism.

$\endgroup$
  • $\begingroup$ What does $\cong$ mean in this context? $\endgroup$ – Omnomnomnom Jun 17 at 1:11
  • $\begingroup$ @Omnomnomnom it is vector bundle isomorphism $\endgroup$ – Ramtin.VA Jun 17 at 1:12
  • 1
    $\begingroup$ This is false. A K3 surface has trivial canonical bundle but not trivial tangent bundle. (en.wikipedia.org/wiki/K3_surface) $\endgroup$ – Gunnar Þór Magnússon Jun 17 at 8:28
  • $\begingroup$ @GunnarÞórMagnússon thanks, it’s wonderful! $\endgroup$ – Ramtin.VA Jun 17 at 9:02
0
$\begingroup$

For a simple way to get counterexamples, note that if $A\cong L_1\oplus\dots\oplus L_n$ splits a direct sum of line bundles, then $\det A\cong L_1\otimes \dots\otimes L_n$. So, for instance, let $X$ be any space with a class $\alpha\in H^2(X,\mathbb{Z})$ such that $\alpha^2\neq 0$ and let $L$ be the line bundle on $X$ with $c_1(L)=\alpha$. Letting $A=L\oplus L^{-1}$ and $B$ be the trivial rank $2$ bundle, then $\det A\cong L\otimes L^{-1}$ and $\det B$ are both trivial, but $A\not\cong B$ since $c_2(A)=c_1(L)c_1(L^{-1})=-\alpha^2\neq 0$ so $A$ is nontrivial.

$\endgroup$
  • $\begingroup$ Thank you very much.It is clear. $\endgroup$ – Ramtin.VA Jun 20 at 6:00

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.