Is it true that, for any $1$-form on a $3$-mfld $\alpha$, $\alpha \wedge d\alpha=0$?
I'd argue like $$0=d(0)=d(\alpha \wedge \alpha)= 2d(\alpha)\wedge \alpha$$
since for any $1$-form $\alpha \wedge \alpha=0$, but then there are counterexample like $\alpha= dz+xdy$ on $\Bbb R^3$ s.t. $\alpha \wedge d\alpha\neq0$
Where am I making a mistake?