Noncommutative unfolding of hypersurface singularity (Q2264074)

An isolated hypersurface singularity is a polynomial \(f\in k[x_1,\dots,x_n]\) for which the Milnor number \(\mu(f)=\dim k[x_1,\dots,x_n]/(\frac{\partial f}{\partial x_1},\dots,\frac{\partial f}{\partial x_n})\) is finite. An unfolding of a hypersurface singularity is a family of hypersurface singularities parametrized by an affine space. This is the problem of deformations of the \(k[y]\)-algebra \(k[x_1,\dots,x_n]\) given by sending \(y\) to \(y=f(x_1,\dots,x_n).\) This article studies noncommutative unfoldings of hypersurface singularities which means deformations of the \(k[y]\)-algebra \(A=k[x_1,\dots,x_n]\) in the category of associative algebras. This can be seen as the study of the Hochschild cochain complex of \(A\) considered as \(k[y]\)-algebra with the Kontsevich Formality theorem as the possible ideal answer. The proof of the formality theorem shows that for a smooth algebra \(A\), there is a weak equivalence between the Hochschild complex to an algebra of polyvector fields. In the case of this article, there is a quasi-isomorphism between \(A\) and a smooth dg algebra over \(k[y]\), and the proof of the formality theorem can be, due to the authors, easily generalized to this setup. Thus the Hochschild cochain complex can be replaced with a certain algebra of polyvector fields which in this case is a dg algebra. The first part of the article proves the following theorem: Let \(\mathbb Q\subset R\) be a commutative ring, and let \(A\) be a commutative smooth dg \(R\)-algebra that is non-positively graded, semifree over \(A^0\) which is smooth as \(R\)-algebra. Then the Hochschild cochain complex of \(A\) over \(R\) is equivalent to the dg algebra of polyvector fields as (homotopy) Gerstenhaber algebras. For a smooth dg algebra \(A\), the algebra of polyvector fields is designed as \(S_A(T[-1])\) with \(T=\text{Der}_R(A,A)\). Because \(S_A(T[-1])\) is a Gerstenhaber algebra, its Harrison chain complex \(B_{\text{Com}^\bot}(S_A(T[-1])\) has structure of a dg Lie bialgebra. A homotopy Gerstenhaber Lie algebra structure on the Hochschild complex \(C(A)\) can be defined via a dg Lie bialgebra structure on \(F^\ast_{\text{Lie}}(C(A)[1])\). An equivalence between Gerstenhaber algebras of polyvector fields \(S_A(T[-1])\) and the Hochschild complex \(C(A)\) is presented on the level of these Lie algebra models. The authors present a dg Lie algebra \(\xi(A)\) and two weak equivalences \(\xi(A)\rightarrow B_{\text{Com}^\bot}(S_A(T[-1])\) and \(\xi(A)\rightarrow F^\ast_{\text{Lie}}(C(A))[1])\) of dg Lie bialgebras. The second weak equivalence is deduced from a dg version of the Hochschild-Kostant-Rosenberg theorem. The setup of dg smooth algebras makes this deduction quite nontrivial and is the main deviation from the proof of Dolgushev-Tamarkin-Tsygan. In this theory, the Poisson bracket appears as a representative of the first order deformation extendable to a second order deformation. Thus the following definition is given: A quasiclassical datum of quantization of a \(B\)-algebra \(A\) is its deformation over \(k[h]/(h^2)\) extendable to \(k[h]/(h^3)\). Thus a quasiclassical datum for a quantization of the ring of smooth functions on a manifold is precisely a Poisson bracket on the manifold. Let \(f\in k[x_1,\dots,x_n]\) be an isolated singular hypersurface and let \(W\) be a vector subspace of \(k[x_a,\dots,x_n]\) complementary to the ideal \((\frac{\partial f}{\partial x_1},\dots,\frac{\partial f}{\partial x_n})\). The authors prove in this article that the quasiclassical data for an MC unfolding of an isolated singularity \(f\in k[x_1,\dots,x_n]\) are given by pairs \((p,S)\) where \(p\in W\) and \(S\) is a Poisson vector field satisfying the condition \([f,S]=0\). With this, the main result of the article is the following: Let \(f\in k[x,y,z]\) define an isolated surface singularity. Then any quasiclassical datum of NC unfolding of \(f\) can be quantized to a noncommutative unfolding over \(k[[h]].\) The proof of the above results involve the the study of a Lie algebroid over a commutative \(k\)-algebra \(A\), such that \(S_A(T[-1])\) has a natural structure of Gerstenhaber algebra. The main thing to solve, is to bypass the fact that the Hochschild complex \(C(A)\) is not a (genuine) Gerstenhaber algebra, this structure is only up to homotopy. The solution comes with Koszul duality and standard homotopy theory for colored operads. The involved subjects are explained through the preliminaries, that is colored operads, Koszul duality, \(\mathcal O_\infty\)-algebras, \(\tilde B\)-algebras. The last gives an algebra structure on the Hochschild cochain complex, and the article contains a section on Lie bialgebras versus \(G\)-algebras (where the operad \(G\) is Koszul). Finally, the different Lie algebra models for the Hochschild cochain complex are classified, that is, their types of equivalence is discussed. For an example, it is proved that the \(\text{Lie}_\infty\)-structure on \(C([1])\) defined by the \(\tilde B\)-structure, coincides with the (strict) Gerstenhaber bracket. Other results proves and explains weak equivalences between the models. Finally the different models and results, in particular the comparison of the different structures on the Hochschild complexes are used to give an application on noncommutative unfolding (of a singularity). This is done by the notion of ``quasiclassical datum'' (defined as deformations of the \(B\)-algebra \(A\) over \(k[h]/(h^2)\) extendable to \(k[h]/(h^3)\)). The article gives nice examples and applications of the noncommutative theory. The stack-theoretic aspect is somehow hidden behind generalities, making the text reasonably easy to read. Also, the results are nice, proven with good techniques.