9
$\begingroup$

Over a ground field of characteristic zero, the algebra of differential operators $\mathcal{Diff}(X)$ on a smooth affine variety $X$ is generated by the functions $O_X$ and the derivations $Der(O_X)$, both viewed as subspaces of the endomorphism algebra $End(O_X)$ of the vector space $O_X$. This is a theorem of Grothendieck that one can find proved in detail in a few places, likethe book Noncommutative Noetherian Rings J. C. McConnell and J. C. Robson or, if I recall correctly, J. E. Bjork's Rings of differential operators (but maybe Bjork does only the analytic case? I don't have the book at hand)

Now there are some obvious relations one knows that hold in $\mathcal{Diff}(X)$ in that case: functions commute, the product of a function and a derivation is given by the module structure of the derivations, the commutator of two derivations is, well, their commutator, which is also a derivation, and the commutator of a derivation and a function is the value of the former on the latter.

Are these relations enough to give a presentation of the algebra?

One could phrase this in the language of Lie-Rinehart pairs: is $\mathcal{Diff}(X)$ the universal enveloping algebra of the Lie—Rinehart pair $(O_X,Der(O_X))$?

The answer to this question is yes, and it is part of what one needs to prove in order to show that the associated graded algebra of $\mathcal{Diff}(X)$ with respect to the filtration on differential order is the coordinate ring of the cotangent bundle on $X$, for example.

Has a detailed proof of this been written down?

$\endgroup$
  • 1
    $\begingroup$ Though it may involve more relations than you want, perhaps you might like EGA IV$_4$ 16.8 (see in particular 16.8.8), 16.10.1 (especially condition (ii) there), and 16.12.2. $\endgroup$ – nfdc23 Dec 11 '17 at 21:50
4
$\begingroup$

A paper by Vladimir Bavula addresses this question. Indeed the relations in question are sufficient to determine $Diff(X)$. Moreover, Bavula gives an explicit finite set of generators of $Diff(X)$ (functions and vector fields) and presents defining relations between the generators.

$\endgroup$

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.