nLab formal scheme

under construction

Idea

A formal scheme is a formal neighbourhood of a scheme, a scheme with infinitesimal thickening.

Generally this may be formalized via ind-objects of schemes. Hence regarded as a sheaf on affine schemes, a formal scheme is a filtered colimit of ordinary schemes (e.g. Strickland 00, section 4).

Regarded as a locally ringed space a formal scheme has the underlying topological space of an ordinary scheme, but its structure sheaf contains extra nilpotent elements.

Motivation

Formal power series rings $k[\![x_1,\ldots,x_n]\!]$ are limits of their truncations (e.g. in one variable $k[\![x]\!]/(x^n)$); they can be viewed as completions and they get equipped with a natural filtration and adic topology.

They do not converge as a series (and make sense) in an open set or in any of the standard topologies (e.g. Zariski and complex topology over $\mathbb{Z}$), but they are rather “localized” in an infinitesimal neighborhood of the origin. One would like to be able to talk about functions supported only infinitesimally (in the transverse direction) to a closed subscheme. The formal schemes of Grothendieck are ringed spaces containing the information on all infinitesimal neighborhoods. Zariski’s theorem on formal functions and establishing the theory of formal groups were some of the concrete motivations.

Definition

There are roughly four equivalent definitions of a $k$-formal scheme for a field $k$ (check this):

Let $Mf_k$ denote the category of finite dimensional $k$-rings (=$k$-algebras which are rings).

1. A $k$-scheme is called a $k$-formal scheme if it is is equivalent to a directed colimit of finite (affine) $k$-schemes.

2. A $k$-scheme is a $k$-formal scheme if it is presented by a profinite $k$-ring; i.e a $k$-ring which is the limit of topologically discrete quotients which are finite $k$-rings. If $A$ is such a topological $k$-ring $Spf_k(A)(R)$ denotes the set of continous morphisms from $A$ to the topologically discrete ring $R$. We have $Spf_k$ is a (contravariant) equivalence between the category of profinite $k$-rings and the category $fSch_k$ of formal $k$-schemes.

3. Instead of defining the category $fSch_k$ of formal $k$-schemes as the opposite of $Mf_k$ define it instead covariantly on the category of finite dimensional $k$-corings.

4. A formal $k$-scheme is precisely a left exact (commuting with finite limits) functor $X:Mf_k\to Set$.

(Demazure, p-divisible groups, chapter I)

The inclusion $Mf_k\hookrightarrow M_k$ induces a functor

$${}^\hat\; :Sch_k\to fSch_k$$

called completion functor.

Formal spectra of Noetherian rings

If $X$ is a scheme, a closed subscheme $Y\subset X$ is given by an embedding of topological spaces with the comorphism $\mathcal{O}_X\to f_*\mathcal{O}_Y$ which is a surjection; but alternatively $\mathcal{O}_Y$ can be recovered from $X$ and the defining sheaf of ideals $\mathcal{I}\subset \mathcal{O}_X$. Then one defines the structure sheaf of the completion $\mathcal{O}_{\hat{X}}$ as $lim_n \mathcal{O}_X/\mathcal{I}^n$ restricted to $Y$ and the completion $\hat{X} := (Y,\mathcal{O}_{\hat{X}})$. If $X = Spec\,R$ where $R$ is a noetherian $I$-adic ring, and $Y=Spec\,R/I$ then the completion is called the formal spectrum of $R$ denoted

$$Spf\,R = \hat{X}$$

(where $R$ is viewed as a topological ring). The formal spectrum is an ind-object in the category of algebraic schemes, viewed as a formal colimit $colim_n Spec (R/I^n)$.

A (locally) noetherian formal scheme is a formal completion of a (locally) noetherian scheme along a closed subscheme. Equivalently, a locally noetherian scheme is a locally ringed space which is locally isomorphic to the formal spectrum of a complete separated adic noetherian ring.

More general ind-schemes

The formal spectrum can be extended to a somewhat bigger class of topological rings than the noetherian ones; Grothendieck developed the theory in the generality of pseudocompact topological rings. However, some important rings, e.g. the ring of integers $\mathbb{Z}$, do not have a pseudocompact topology. Thus one could try to consider a more general subcategory of ind-schemes (with at least the requirement that the ind-object be represented by a diagram where the connecting morphisms are closed immersions of schemes); one such approach is outlined in some detail in (Beilinson-Drinfeld).

Examples

Formal power series and their formal spectra

The completion of a ring of the polynomial ring $\mathbb{Z}[x]$ (the formal dual to the affine line $\mathbb{A}^1$) at the ideal $(x)$ is the limit – formed in the category CRing –

$$\mathbb{Z}[x]^\wedge_{(x)} = \underset{\longleftarrow}{lim}_n (\mathbb{Z}[x])/(x^{n+1})$$

of the quotient rings of $\mathbb{Z}[x]$ by the ideals generated by $x^{n+1}$ (the ring of dual numbers and its higher analogs). This yields the formal power series ring

$$\mathbb{Z}[x]^\wedge_{(x)} \simeq \mathbb{Z}[ [x] ] \,.$$

Notice that the formal power series ring contains no nilpotent elements except for zero – even though each filtering stage $(\mathbb{Z}[x])/(x^n)$ does.

In contrast, the filtered colimit of the sheaves on affine varieties $Spec(R)$ that are represented by this system of nilpotent rings is the formal scheme

$$\widehat{\mathbb{A}^1} = \underset{\longrightarrow}{lim}_n Spec(\mathbb{Z}[x]/(x^{n+1})) \;\;\; \in Sh(Aff) = Sh(CRing^{op})$$

given by the sheaf which sends each ring to its nilradical

$$\widehat{\mathbb{A}^1} \;\colon\; Spec(R) \mapsto Nil(R) \,.$$

(see e.g. Strickland 00, example 4.2, example 4.18).

Formal neighbourhood of the diagonal

See at formal neighbourhood of the diagonal.

Formal groups

The group objects in formal schemes are the formal groups. See there for more.

References

Classical discussion includes

• Alexander Grothendieck, Géométrie formelle et géométrie algébrique, FGA 2 (Séminaire Bourbaki, t. 11, 1958/59, no. 182)

• Luc Illusie, Grothendieck existence theorem in formal geometry, chapter 8 in Fantechi, Gottsche, Illusie, Kleiman, Nitsure, Vistoli, Fundamental algebraic geometry, Grothendieck’s FGA explained, Math. Surveys and Monographs 123, AMS 2005 (draft version pdf)

• A. Grothendieck (avec J. Dieudonne), EGA I.10

• A. Grothendieck et al. SGA III.2 (Exp. 7a, P. Gabriel, Étude infinitésimale des schémas en groupe et groupes formels; Exp. 7b, P. Gabriel, Groupes formels)

• R. Hartshorne, Algebraic geometry, II.9

• M. Demazure, P. Gabriel, Groupes algebriques, tome 1 (later volumes never appeared), Mason and Cie, Paris 1970

• Michel Demazure, lectures on p-divisible groups web

• Daniel Murfet, notes Section2.9-FormalSchemes.pdf

• Leovigildo Alonso, Ana Jeremias, Marta Perez, Infinitesimal lifting and Jacobi criterion for smoothness on formal schemes (arXiv:math/0604241)

More general discussion in terms of ind-schemes includes

• A. Beilinson, V. Drinfel’d, Quantization of Hitchin’s integrable system and Hecke eigensheaves on Hitchin system, preliminary version (pdf)

Another approach using a certain topological extension of the Yoneda lemma on $k Alg^{op}$ has been proposed in

• B. Pareigis, R. A. Morris, Formal groups and Hopf algebras over discrete rings, Trans. Amer. Math. Soc. 197 (1974), 113–129 (doi, nlab entry).

Nikolai Durov has considered a flexible bigger category (which inludes the usual schemes) of covariant functors from the category of pairs $(R,I)$ where $R$ is a commutative ring and $I$ a nilpotent ideal; the correspondence between formal groups and Lie algebras based on Hausdorff series is neatly developed and used in that language; see chapters 7–9 of

• N. Durov, S. Meljanac, A. Samsarov, Z. Škoda, A universal formula for representing Lie algebra generators as formal power series with coefficients in the Weyl algebra, Journal of Algebra 309, n. 1, 318–359 (2007) (doi:jalgebra) (math.RT/0604096).

For another generalization of formal schemes see

• T. Yasuda, Non-adic formal schemes, Int. Math. Research Notices 2009: 2417–2475, (doi:imrn, arxiv)

In a fundamental article in noncommutative algebraic geometry,

• M. Kapranov, Noncommutative geometry based on commutator expansions, math.AG/9802041,

Kapranov introduced objects which should be interpreted as the infinitesimal neighborhoods of those commutative schemes with a closed immersion into a noncommutative scheme which is locally isomorphic to the spectrum of a free associative algebra.

And a decidedly functorial perspective is in (section 4 of)

• Neil Strickland, Formal schemes and formal groups, math.AT/0011121

For the scheme geometric picture behind the infinitesimal neighborhoods and D-modules see also

• A. Beĭlinson, J. Bernstein, J., A proof of Jantzen conjectures, in I. M. Gelʹfand Seminar, 1–50, Adv. Soviet Math., 16, Part 1, Amer. Math. Soc., Providence, RI, 1993 (MR1237825 (95a:22022))

Some aspects of formal completions from the point of view of the derived categories are in

• D. Orlov, Formal completions and idempotent completions of triangulated categories of singularities, arxiv/0901.1859

• Alexander I. Efimov, Formal completion of a category along a subcategory, arxiv/1006.4721