Алгебры, заданные потенциалом, и 3-Калаби-Яу алгебры

In his 2007 paper Calabi-Yau Algebras [Gin07], Victor Ginzburg introduces the eponymous algebras in order to `transplant most of conventional Calabi- Yau geometry’ to the noncommutative setting. Indeed, as the author argues, these algebras appear naturally in the study of Calabi-Yau (CY) manifolds and the homological properties that de ne them arise directly from the ge- ometry of the associated spaces.
Though the de nition of a CY algebra makes sense for all (associative) al- gebras, we will restrict ourselves to the homogeneous N-graded connected (A0 = k) case. Of particular interest amongst those are 3-CY algebras, which were shown to be de ned by potentials [Boc08, Theorem 3.1]. This result immediately yields the following question.
Problem. When does a homogeneous potential give rise to a 3-Calabi-Yau algebra?
In fact, this problem already appeared in Ginzburg’s seminal paper where the author pointed that `there seems to be no simple characterisation of those potentials’. Since then, some work has been done to better understand this question, notably classi cations providing many examples and non-examples (e.g. [IS18] and [MS17] for cubic algebras on two generators and quadratic algebras on three generators, including twisted ones).
In addition to having a nice description via potentials, 3-Calabi-Yau algebras enjoy close links with neighbouring homological notions. The most important is the Koszul condition, generalised to arbitrary degree in [Ber01]. It is much weaker than the Calabi-Yau condition, and many examples of Koszul non-CY algebras exist (see for instance Example 8).
It was however shown by Berger and Solotar [BS13] that 3-Calabi-Yau alge- bras are exactly those Koszul algebras which enjoy an additional property, called here proper. Their result can be formulated in terms of Hilbert series (see Theorem 11 below), which can be computed using Groebner bases, and is hence easier to check than more abstract homological properties.
The aim of the present work will be to investigate whether certain potentials are proper (the additional property mentioned above). Because the goal is to reduce the question of whether an algebra is 3-Calabi-Yau to whether it is Koszul, and 3-CY are domains, we only consider domains (see Example 9 for a proper non-domain).
After recalling de nitions and important results in section 2, we prove in section 3 that a homogeneous twisted domain on two or three generators is proper.