The deformation theory of sheaves of commutative rings (Q435934)

If \(f:X\to Y\) is a smooth morphism of schemes, and \(Y'\) is a square-zero extension of \(Y\) such that the ideal of \(Y\) in \(Y'\) is \(\mathcal O_Y\), then the square-zero extensions \(X'\) of \(X\) over \(Y\) are obstructed by a class \(\omega\in H^2(X,T_{X/Y})\), where \(T_{X/Y}\) is the relative tangent bundle. Should this class vanish, deformations form a torsor under \(H^1(X,T_{X/Y})\) and automorphisms of any fixed deformation are in bijection with \(H^0(X,T_{X/Y}).\) If \(f\) is non-smooth, a similar description of obstructions, deformations and automorphisms persists: Letting \(\mathbf{L}_{X/Y}\) denote the cotangent complex of \(X\) over \(Y\), there is an obstruction \(\omega\in \text{Ext}^2(\mathbf{L}_{X/Y},\mathcal O_X)\) whose vanishing is equivalent to the existence of a deformation. If a deformation exists, the deformations form a torsor under \(\text{Ext}^1(\mathbf{L}_{X/Y},\mathcal O_X)\) and automorphisms of any given solution are in bijection with \(\text{Ext}^0(\mathbf{L}_{X/Y},\mathcal O_X)\). The failure of the ``gerbe argument'' in the non-smooth case can be attributed to the fact that the deformation problem is not locally trivial in the Zariski topology on \(X\). To interpret Illusie's result in terms of gerbes, a topology must be found that is finer than the Zariski topology, and such that the deformation problem becomes locally trivial in this topology. In the affine situation, such a topology was found independently by Quillen and Rim, and Quillen generalized this to schemes in general without publishing the result. Gaitsgory defined a topology on the category of associative algebras on a scheme and showed that it is fine enough to find local trivializations of deformation problems associated to quasi-coherent algebras. His methods can be applied to the commutative case, where they can be used to treat the relative deformation theory of a scheme that is affine over the base. If \(f:X\to Y\) is a morphism of schemes, then \(\mathcal O_X\) is in general not quasi-coherent as an \(f^{-1}\mathcal O_Y\)-algebra, so Gaitsgory's results do not apply directly to the deformation of schemes. The introduction of banded gerbes to explain the obstruction to the existence of algebra extensions is due to Gaitsgory. That deformations, when they exists, can be viewed as torsors was observed by Quillen. In this article, Wise defines a new topology on the category of all commutative rings in a topos and show that it is fine enough to to trivialize the standard deformation problems for commutative rings, but still coarse enough to glue their solution. Cohomological obstructions to the existence of solutions are given, and solutions to the problems is found when they vanish. This is used to apply the ideas of Quillen, Rim and Gaitsgory to the deformation theory of schemes. The approach to cotangent cohomology given here is compared to Illusie's, and it is proved that the author's obstruction groups agrees with his. The main results of the article is as follows: Suppose that \(f:X\to Y\) is a morphism of \(Z\)-schemes, and consider consider the problem of extending \(f\) to a fixed square-zero extension \(X'\) of \(X\) over \(Z\) with ideal \(J\). A site \(g^{-1}\mathcal O_Z\text{-}\underline{\text{Alg}}/f^{-1}\mathcal O_Y\) abbreviated \(\mathcal O_Z\text{-}\underline{\text{Alg}}/\mathcal O_Y\) is defined and proved to be fine enough to ensure that the deformation problem i locally trivial. The first main result is that the extensions \(\text{Hom}^X(X',Y)\) of \(f\) to \(f':X'\to Y\) form a torsor on \(\mathcal O_Z\text{-}\underline{\text{Alg}}/\mathcal O_Y\) under the sheaf of abelian groups \(\underline{\text{Der}}_{\mathcal O_Z}(\mathcal O_Y,J)\). The class of this torsor in \(H^1(\mathcal O_Z\text{-}\underline{\text{Alg}}/\mathcal O_Y,\underline{\text{Der}}_{\mathcal O_Z}(\mathcal O_Y,J))\) obstructs the existence of a lift. Provided that the obstruction vanishes, all lifts form a torsor under \(\text{Der}_{\mathcal O_Z}(\mathcal O_Y,J)\). Consider a morphism of schemes \(f:X\to Y\) and a fixed extension \(Y'\) of \(Y\) with ideal \(I\). Given a homomorphism \(\phi:f^\ast I\to J\) for some quasi-coherent sheaf \(J\) on \(X\). Then it is searched for a completion of the diagram \[ \begin{tikzcd} X \rar\dar & X^\prime\dar\\ Y & Y^\prime \end{tikzcd} \] in which \(X'\) is a square-zero extension of \(X\) by the ideal \(J\) and the induced morphism \(f'{}^\ast I\to J\) coincides with \(\phi\). This deformation problem also becomes locally trivial in \(f^{-1}\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X:=\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X\). Then the completions of the diagram form a gerbe on \(\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X\) banded by \(\underline{\text{Der}}_{\mathcal O_Y}(\mathcal O_X,J)\). The class in \(H^2(\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X,\underline{\text{Der}}_{\mathcal O_Y}(\mathcal O_X,J))\) of this gerbe obstructs the existence of a solution to this problem. If the obstruction vanishes, solutions form a torsor under \(H^1(\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X,\underline{\text{Der}}_{\mathcal O_Y}(\mathcal O_X,J))\), and the automorphism group of any solution is \(H^0(\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X,\underline{\text{Der}}_{\mathcal O_Y}(\mathcal O_X,J)).\) In the case where \(I=0\) and \(Y=Y'\), an extension always exists, so this implies that there is an equivalence of categories between the category of completions of the diagram \[ \begin{tikzcd} X \ar[rr] \ar[dr] & & X^\prime\ar[dl]\\ & Y &\end{tikzcd} \] by a scheme \(X'\) that is a square-zero extension of \(X\) with ideal \(J\) and the category of torsors on \(\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X\) under the sheaf of abelian groups \(\underline{\text{Der}}_{\mathcal O_Y}(\mathcal O_X,J)\). Isomorphism classes are in bijection with \(H^1(\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X,\underline{\text{Der}}_{\mathcal O_Y}(\mathcal O_X,J))\) and automorphisms of any object are in bijection with \(H^0(\mathcal O_Y\text{-}\underline{\text{Alg}}/\mathcal O_X,\underline{\text{Der}}_{\mathcal O_Y}(\mathcal O_X,J)).\) The final result then proves that the obstruction groups are the same as Illusie's. The article contains the necessary definitions of stacks and gerbes so that the problem can be stated. Then the job of making a deformation theory for all rings in a topos is done in a nice way. The results are necessary for further studies, and the work done had to be made.