Archive for October 2024

And Coherent Sheaves...

In the last post, we discussed the quasi-coherent sheaves for some ring $R$ and scheme $X$. On a scheme $X$, the quasi-coherent sheaves form an abelian category and in fact, this category is a sub-category of the category of $R$-modules. So, simply, as Vakil puts it in his notes, one should better look if the category of $R$-modules is an abelian category and prove that the category of quasi-coherent sheaves (call $Q_{coh}$) is indeed a subcategory of the category of $R-$modules ($Mod_R$), so

$$Q_{coh} \subset Mod_R$$
I will leave this to you to prove this.

Similarly, the coherent sheaves for some ring $R$ also form the abelian category (similar proof) and in fact, quasi-coherent sheaves will not always form an abelian category for any arbitrary ringed space while coherent sheaves will always form an abelian category. Coherent sheaves come with a bit more than quasi-coherent sheaves, both attached very strongly to the sheaf of modules of $R$. The extra condition is of finite presentation and finitely generated modules. A quasi-coherent sheaf is a coherent sheaf if the modules $M$ are finitely generated (hence $R^n \to M$ which is a surjection). For the Noetherian scheme, a finitely generated quasi-coherent sheaf will automatically be a coherent sheaf. But for non-Noetherian schemes, it is not guaranteed. That is why, one should be careful defining coherent sheaf as quasi-coherent sheaf which is finitely generated, which is not true always.

In general, local nature of the ringed space will be described better by the coherent sheaf (category). The discussion on coherent sheaf and finite presentation (that I do not discuss in this) will be done later. Some resources on this subject are this, this by Serre and Vakil's notes.

Posted in | Leave a comment Print it.