Abstract
In 1988 Simpson extended the Donaldson–Uhlenbeck–Yau theorem to the context of Higgs bundles, and as an application he proved a uniformization theorem which characterizes complex projective manifolds and quasi-projective curves whose universal coverings are complex unit balls. In this paper we give a necessary and sufficient condition for quasi-projective manifolds to be uniformized by complex unit balls. This generalizes the uniformization theorem by Simpson. Several byproducts are also obtained in this paper.
Similar content being viewed by others
Data availability
All data generated or analysed during this study are included in this article.
Notes
References
Arbarello, E., Harris, J.D., Cornalba, M., Griffiths, P.: Geometry of Algebraic Curves: Volume ii with a Contribution By Joseph Daniel Harris, Grundlehren der mathematischen Wissenschaften. Springer, Berlin (2011)
Ash, A., Mumford, D., Rapoport, M., Tai, Y.-S.: Smooth Compactifications of Locally Symmetric Varieties. With the Collaboration of Peter Scholze, 2nd edn. Cambridge University Press, Cambridge (2010).. (English)
Baily, W.L., Borel, A.: Compactification of arithmetic quotients of bounded symmetric domains. Ann. Math. 84, 442–528 (1966)
Boucksom, S., Eyssidieux, P., Guedj, V., Zeriahi, A.: Monge–Ampère equations in big cohomology classes. Acta Math. 205(2), 199–262 (2010). (English)
Biquard, O.: Fibrés de Higgs et connexions intégrables: Le cas logarithmique (diviseur lisse). Ann. Sci. Éc. Norm. Supér. (4) 30(1), 41–96 (1997). (French)
Borel, A.: Compact Clifford–Klein forms of symmetric spaces. Topology 2, 111–122 (1963)
Borel, A.: Introduction aux groupes arithmétiques, Actualités scientifiques et industrielles. Hermann, Paris (1969)
Boucksom, S.: On the volume of a line bundle. Int. J. Math. 13(10), 1043–1063 (2002). (English)
Brunebarbe, Y.: Semi-positivity from Higgs bundles, arXiv e-prints. arXiv:1707.08495 (2017)
Baldi, G., Ullmo, E.: Special subvarieties of non-arithmetic ball quotients and Hodge theory. arXiv e-prints. arXiv:2005.03524 (2020)
Cadorel, B.: Symmetric differentials on complex hyperbolic manifolds with cusps. J. Differ. Geom. 118(3), 373–398 (2021)
Cornalba, M., Griffiths, P.: Analytic cycles and vector bundles on non-compact algebraic varieties. Invent. Math. 28, 1–106 (1975). (English)
Chavel, I.: Riemannian Geometry. A Modern Introduction, vol. 78, 2nd edn. Cambridge University Press, Cambridge (2006).. (English)
Cattani, E., Kaplan, A., Schmid, W.: Degeneration of Hodge structures. Ann. Math. (2) 123(3), 457–535 (1986)
Cheng, S.-Y., Yau, S.-T.: On the existence of a complete Kähler metric on non-compact complex manifolds and the regularity of Fefferman’s equation. Commun. Pure Appl. Math. 33, 507–544 (1980). ((English))
Cheng, S.-Y., Yau, S.-T.: Inequality between Chern numbers of singular Kähler surfaces and characterization of orbit space of discrete group of SU(2,1). Complex differential geometry and nonlinear differential equations. In: Proceedings of AMS-IMS-SIAM Joint Summer Research Conference, Brunswick/Maine 1984, Contemporary of Mathematics, vol. 49, pp. 31-44 (1986)
Debarre, O.: Higher-Dimensional Algebraic Geometry. Springer, New York (2001).. (English)
Demailly, J.-P.: Analytic Methods in Algebraic Geometry, vol. 1. International Press, Somerville; Higher Education Press, Beijing (2012) (English)
Demailly, J.-P.: Complex analytic and differential geometry, online book (2012). https://www-fourier.ujf-grenoble.fr/~demailly/manuscripts/agbook.pdf
Deng, Y., Hao, F.: Vanishing theorem for tame harmonic bundles via \(L^2\)-cohomology. arXiv e-prints. arXiv:1912.02586 (2019)
Eyssidieux, P.: Orbifold Kähler Groups related to arithmetic complex hyperbolic lattices. arXiv:1805.00767 (2018)
Greb, D., Kebekus, S., Peternell, T., Taji, B.: Harmonic metrics on Higgs sheaves and uniformization of varieties of general type. Math. Ann. 378, 1061–1094 (2020)
Greb, D., Kebekus, S., Peternell, T., Taji, B.: The Miyaoka–Yau inequality and uniformisation of canonical models. Ann. Sci. Éc. Norm. Supér. (4) 52(6), 1487–1535 (2019). (English)
Greb, D., Kebekus, S., Taji, B.: Uniformisation of higher-dimensional minimal varieties. arXiv e-prints. arXiv:1608.06644 (2016)
Guenancia, H., Taji, B.: Orbifold stability and Miyaoka–Yau inequality for minimal pairs. arXiv e-prints. arXiv:1611.05981 (2016)
Guenancia, H.: Kähler–Einstein metrics with mixed Poincaré and cone singularities along a normal crossing divisor. Ann. Inst. Fourier 64(3), 1291–1330 (2014). (English)
Guenancia, H.: Semistability of the tangent sheaf of singular varieties. Algebraic Geom. 3(5), 508–542 (2016). (English)
Harris, M.: Functorial properties of toroidal compactifications of locally symmetric varieties. Proc. Lond. Math. Soc. 1, 1–22 (1989)
Hacon, C.D., McKernan, J.: On Shokurov’s rational connectedness conjecture. Duke Math. J. 138(1), 119–136 (2007). (English)
Kazhdan, D.: On arithmetic varieties. II. Isr. J. Math. 44, 139–159 (1983). ((English))
Kollár, J., Mori, S.: Birational Geometry of Algebraic Varieties. With the Collaboration of C. H. Clemens and A. Corti, vol. 134. Cambridge University Press, Cambridge (1998).. (English)
Kobayashi, R., Nakamura, S., Sakai, F.: A numerical characterization of ball quotients for normal surfaces with branch loci. Proc. Jpn. Acad. Ser. A 65(7), 238–241 (1989). ((English))
Kobayashi, R.: Einstein–Kähler metrics on open algebraic surfaces of general type. Tohoku Math. J. (2) 37, 43–77 (1984). (English)
Kobayashi, R.: Einstein–Kähler V-metrics on open Satake V-surfaces with isolated quotient singularities. Math. Ann. 272, 385–398 (1985). (English)
Kobayashi, R.: Uniformization of complex surfaces, Kähler metric and moduli spaces. Adv. Stud. Pure Math. 18(2), 313–394 (1990)
Kollár, J.: Shafarevich maps and plurigenera of algebraic varieties. Invent. Math. 113, 177–215 (1993)
Kollár, J.: Lectures on Resolution of Singularities (AM-166). Princeton University Press, Princeton (2007)
Maubon, J.: Higgs bundles and representations of complex hyperbolic lattices, Handbook of group actions (vol. II). In: Ji, L., Papadopoulos, A., Yau, S.-T. (eds.) Advanced Lectures in Mathematics, vol. 32, pp. 201–244. International Press, Somerville; Higher Education Press, Beijing (2015)
Mochizuki, T.: Asymptotic behaviour of tame nilpotent harmonic bundles with trivial parabolic structure. J. Differ. Geom. 62(3), 351–559 (2002)
Mochizuki, T.: Kobayashi–Hitchin correspondence for tame harmonic bundles and an application. Astérisque 309, viii+117 (2006)
Mochizuki, T.: Asymptotic behaviour of tame harmonic bundles and an application to pure twistor D-modules. I. Mem. Am. Math. Soc. 185(869), xii+324 (2007)
Mochizuki, T.: Good wild harmonic bundles and good filtered Higgs bundles. arXiv e-prints. arXiv:1902.08298v3 (2019)
Mok, N.: Metric Rigidity Theorems on Hermitian Locally Symmetric Manifolds, vol. 6. World Scientific, Singapore (1989).. (English)
Mok, N.: Projective algebraicity of minimal compactifications of complex-hyperbolic space forms of finite volume. Perspectives in analysis, geometry, and topology. On the occasion of the 60th birthday of Oleg Viro. Based on the Marcus Wallenberg symposium on perspectives in analysis, geometry, and topology, Stockholm, Sweden, May 19–25, 2008, pp. 331–354. Basel: Birkhäuser (2012) (English)
Mumford, D.: Hirzebruch’s proportionality theorem in the non-compact case. Invent. Math. 42, 239–272 (1977). (English)
Mok, N., Yeung, S.-K.: Geometric realizations of uniformization of conjugates of Hermitian locally symmetric manifolds. Complex analysis and geometry, pp. 253–270. Plenum Press, New York (1993).. (English)
Nadel, A.M.: The nonexistence of certain level structures on abelian varieties over complex function fields. Ann. Math. (2) 129(1), 161–178 (1989). (English)
Noguchi, J., Ochiai, T.: Geometric Function Theory in Several Complex Variables. Transl. from the Japanese by Junjiro Noguchi, vol. 80. American Mathematical Society, Providence (1990).. (English)
Păun, M., Takayama, S.: Positivity of twisted relative pluricanonical bundles and their direct images. J. Algebraic Geom. 27(2), 211–272 (2018). (English)
Raghunathan, M.S.: Discrete Subgroups of Lie Groups, Results and Problems in Cell Differentiation. Springer, Berlin (1972)
Simpson, C.T.: Constructing variations of Hodge structure using Yang–Mills theory and applications to uniformization. J. Am. Math. Soc. 1(4), 867–918 (1988)
Simpson, C.T.: Harmonic bundles on noncompact curves. J. Am. Math. Soc. 3(3), 713–770 (1990)
Simpson, C.T.: Higgs bundles and local systems. Inst. Hautes Études Sci. Publ. Math. 75, 5–95 (1992)
To, W.-K.: Total geodesy of proper holomorphic immersions between complex hyperbolic space forms of finite volume. Math. Ann. 297(1), 59–84 (1993). (English)
Tsuji, H.: A characterization of ball quotients with smooth boundary. Duke Math. J. 57(2), 537–553 (1988). (English)
Tian, G., Yau, S.-T.: Existence of Kähler–Einstein metrics on complete Kähler manifolds and their applications to algebraic geometry. Mathematical aspects of string theory. In: Proceedings of the Conference, San Diego/California 1986, Adavnced Series in Mathematical Physics, vol. 1, pp. 574–629 (1987)
Yau, S.-T.: On the Ricci curvature of a compact Kähler manifold and the complex Monge–Ampère equation. I. Commun. Pure Appl. Math. 31, 339–411 (1978). (English)
Yau, S.-T.: A splitting theorem and an algebraic geometric characterization of locally Hermitian symmetric spaces. Commun. Anal. Geom. 1(3), 473–486 (1993). (English)
Zhang, Y.: Miyaoka–Yau inequality for minimal projective manifolds of general type. Proc. Am. Math. Soc. 137(8), 2749–2754 (2009). (English)
Zuo, K.: On the negativity of kernels of Kodaira–Spencer maps on Hodge bundles and applications. Asian J. Math. 4(1), 279–301 (2000). (English)
Acknowledgements
This work owes a lot to the deep work [40, 51,52,53], to which I express my deepest gratitude. I sincerely thank Professor Carlos Simpson for answering my questions, as well as his suggestions and encouragements. I thank Professor Takuro Mochizuki for sending me his personal notes on the proof of Theorem 2.9. I also thank Professors Jean-Pierre Demailly, Henri Guenancia, Emmanuel Ullmo, Shing-Tung Yau, and Gregorio Baldi, Jiaming Chen, Chen Jiang, Jie Liu, Mingchen Xia for very helpful discussions and their remarks on this paper. My special thanks go to Benoît Cadorel for his very fruitful discussions on the toroidal compactification, which lead to a joint appendix with him in this paper. Last but not least, I am grateful to the referee for his/her careful readings and very helpful comments to improve this manuscript.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Ngaiming Mok.
Appendix written by both Ya Deng and Benoît Cadorel.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work is supported by “le fond Chern” à l’IHES.
Appendix A: Metric rigidity for toroidal compactification of non-compact ball quotients
Appendix A: Metric rigidity for toroidal compactification of non-compact ball quotients
The main motivation of this appendix is to provide one building block for Theorem 5.7.(ii). Our main result, Theorem A.7, says that there is no other smooth compactification for non-compact ball quotient than the toroidal one, so that the Bergman metric grows “mildly” near the boundary. Besides its own interests, this result is applied to show that the smoothness of D in Theorem A is necessary if one would like to characterize non-compact ball quotients.
1.1 A.1: Toroidal compactifications of quotients by non-neat lattices
In this section, we recall a well known way of constructing the toroidal compactifications of ball quotients in the case where the lattice has torsion at infinity. The reader will find more details about the natural orbifold structure on these compactifications in [21]. For our purposes, the basic result given in Proposition A.1 will be sufficient.
Recall that we say that a lattice \(\Gamma \subset PU(n,1)\) is neat (cf. [7]) if for any \(g \in \Gamma \), the subgroup of \({\mathbb {C}}^*\) generated by the eigenvalues of g is torsion free. This implies that \(\Gamma \) is torsion free and that all parabolic elements of \(\Gamma \) are unipotent, so that the toroidal compactifications of \(\frac{{\mathbb {B}}^n}{\Gamma }\) provided by [2, 44] are smooth (there is no "torsion at infinity"). Note that by [7, Proposition 17.4] in the arithmetic case, and [6], or [50, Theorem 6.11] in general, any lattice in PU(n, 1) admits a finite index neat normal sublattice.
Proposition A.1
Let \(\Gamma \subset PU(n, 1)\) be a torsion free lattice, and let \(\Gamma ' \subset \Gamma \) be a finite index normal neat sublattice. Let \(U = \frac{{\mathbb {B}}^n}{\Gamma }\), \(U' = \frac{{\mathbb {B}}^n}{\Gamma '}\), and denote by \(X'\) the smooth toroidal compactification of \(U' = \frac{{\mathbb {B}}^n}{\Gamma '}\) as constructed in [2, 44].
Then the natural action of the finite group \(G = \frac{\Gamma }{\Gamma '}\) on \(U'\) extends to \(X'\), and the quotient \(X = \frac{X'}{G}\) is a normal projective space, with boundary \(X - U\) made of quotient of abelian varieties by finite groups. Moreover, when \(\Gamma \) is arithmetic, X coincides with the toroidal compactification of U constructed in [2].
Before explaining how to prove Proposition A.1, let us recall the construction of \(X'\) as it is defined in [44] (see also [11] for a similar discussion).
Each component D of \(X' - U'\) is associated to a certain \(\Gamma '\)-orbit of points of \(\partial {\mathbb {B}}^n\), whose points are called the \(\Gamma '\)-rational boundary components of \(\partial B^n\) (cf. [2, Chapter 3] or [44, §1.3]). Let \(b \in \partial {\mathbb {B}}^n\) be such a point, and let \(N_b \subset PU(n,1)\) be its stabilizer. This is a maximal parabolic real subgroup of PU(n, 1); let us denote by \(W_b\) its unipotent radical. This group is an extension \(1 \rightarrow U_b \rightarrow W_b \overset{\pi }{\rightarrow } A_b \rightarrow 1\), where \(A_b \cong {\mathbb {C}}^{n-1}\), and \(U_b \cong {\mathbb {R}}\) is the center of \(W_b\). Let \(L_b = \frac{N_b}{W_b}\). This reductive group can be embedded as a Levi subgroup in \(N_b\), so that \(N_b = W_b \cdot L_b\). Moreover, we have a further decomposition \(L_b = U(n-1) \times {\mathbb {R}}\). (all this description can be obtained e.g. by specializing the discussion of [3, Section 1.3] or [2, Section 4.2] to the case of the ball).
This Lie theoretic description of \(N_b\) can be understood more easily by expressing the action of the previous groups on the horoballs tangent to b. Let \((S_b^{(N)})_{N \ge 0}\) be the family of these horoballs. Each \(S_b^{(N)} \subset {\mathbb {B}}^n\) can be described as an open subset in a Siegel domain of the third kind, as follows:
We have \(S_b^{(0)} \cong {\mathbb {B}}^n\), and when \(b = (0, ..., 0, 1)\), the change of coordinates between the two descriptions of the ball is given by the Cayley transform
Now, if \(g \in W_b\), we can write \(g = (s, a)\) accordingly to the decomposition \(W_b \overset{sets}{=} U_b \times A_b\) (\(U_b \cong {\mathbb {R}}\), \(A_b \cong {\mathbb {C}}^{n-1}\)), and we have, for any \((w', w_n) \in S_b^{(N)}\):
We check easily that \(S^{(N)}_b\) is preserved by \(W_b\). Also, for any \(g \in L_b \simeq U(n-1) \times {\mathbb {R}}\), we can write \(g = (r, t)\), and we then have
We are now ready to describe the quotients of \(S_b^{(N)}\) by the action of \(\Gamma ' \cap N_b\). Note first that since \(\Gamma '\) is neat, we have \(\Gamma ' \cap N_b \subset W_b\). Then, by the discussion above, we obtain a decomposition as sets \(N_b \overset{sets}{=} ({\mathbb {C}}^{n-1} \times {\mathbb {R}}) \times (U(n-1) \times {\mathbb {R}})\), in which the elements of \(\Gamma ' \cap N_b\) can be written as \((a, t, \mathrm{Id}, 0)\). It also follows from [44] that \(\Gamma ' \cap U_b = {\mathbb {Z}} \tau \) for some \(\tau \in U_b \simeq {\mathbb {R}}\). This last fact permits to form the quotient \(G_b^{(N)} = \frac{S_b^{(N)}}{U_b \cap \Gamma '}\); using (A.1.1), we can also express the latter quotient as an open subset of \({\mathbb {C}}^{n-1} \times {\mathbb {C}}^*\):
and the quotient is then realized by the map \((z', z_n) \in S_b^{(N)} \rightarrow (z', e^{\frac{2i\pi }{\tau } z_n}) \in G_b^{(N)}\).
The group \(\Lambda _b := \pi (\Gamma ' \cap W_b) \subset {\mathbb {C}}^{n-1}\) is an abelian lattice, acting on \(G_b^{(N)} \subset {\mathbb {C}}^{n-1} \times {\mathbb {C}}^*\) as
Clearly, the closure \(\overline{G_b^{(N)}}\) in \({\mathbb {C}}^n\) is an open neighborhood of \({\mathbb {C}}^{n-1} \times \{0\}\). We can form the quotient
which is then isomorphic to a tubular neighborhood of the abelian variety \(\frac{{\mathbb {C}}^{n-1}}{\Lambda _b}\) in some negative line bundle. Finally, the toroidal compactification \(X'\) can be obtained by glueing the open varieties \(\Omega _b^{(N)}\) to \(U'\) (as b runs among a system of representatives of the rational boundary components, and N is large enough).
Our claims about X can be derived from the following lemma.
Lemma A.2
Let \(b \in \partial {\mathbb {B}}^n\) be a \(\Gamma '\)-rational boundary component, and let \(g \in \Gamma \). Then the point \(b' = g \cdot b\) is also \(\Gamma '\)-rational, and there exists \(N, N' > 0\), for which g induces an isomorphism \(S_b^{(N)} \overset{g}{\rightarrow } S_{b'}^{(N')}\), yielding in turn a unique compatible biholomorphism \(\Omega _b^{(N)} \rightarrow \Omega _{b'}^{(N')}\).
Proof
As \(\Gamma '\) is torsion free, a point \(z \in \partial {\mathbb {B}}^n\) is \(\Gamma '\)-rational if and only if \(W_b \cap \Gamma ' \ne \{ e \}\) (see [44, §1.3]). Since g normalizes \(\Gamma '\), we have \(g(W_b \cap \Gamma ')g^{-1} \subset W_{b'} \cap \Gamma '\) so \(b'\) is \(\Gamma '\)-rational if b is.
As for our second claim, since the set of horoballs is preserved by the action of PU(n, 1), we may find \(N, N'\) such that g induces a isomorphism \(S_b^{(N)} \rightarrow S_{b'}^{(N)}\). Let \((x', x_n)\) (resp. \((y', y_n)\)) be standard coordinates on \(S_b^{(N)}\) (resp. \(S_b^{(N')}\)) as in (A.1.1), chosen so that \((y', y_n) = (x', x_n) \circ u\) for some \(u \in U(n)\) satisfying \(u \cdot b' = b\). Then \(ug \in N_b\), and by (A.1.2) and (A.1.3), we have \((y', y_n) \circ g = f(x', x_n)\) for some affine map f.
Since g normalizes \(\Gamma '\), we have \(g (\Gamma ' \cap U_b) g^{-1} = \Gamma ' \cap U_{b'}\), so the map \(S_b^{(N)} \overset{g}{\rightarrow } S_{b'}^{(N')}\) passes to the quotient to give a map \({\widetilde{g}} : G_b^{(N)} {\rightarrow } G_{b'}^{(N')}\). Using an explicit expression for the affine map f, we find an (a priori multivaluate) expression for \({\widetilde{g}}\) as
for some \(A \in M_{n-1}({\mathbb {C}})\), some vectors \(u, b, z_0' \in {\mathbb {C}}^{n-1}\) and \(C, a \in {\mathbb {C}}\). The formula above induces a well-defined, invertible map \(G_b^{(N)} \rightarrow G_{b'}^{(N')}\), so we have \(u=0, a=\pm 1\). This implies that \({\widetilde{g}}\) extends holomorphically to \({\widetilde{g}} : \overline{G_b^{(N)}} \rightarrow \overline{G_{b'}^{(N')}}\). Finally, as g normalizes \(\Gamma '\), \({\widetilde{g}}\) passes to the quotient by \(\Lambda _{b}\) and \(\Lambda _{b'}\), giving a biholomorphism \(\Omega _b^{(N)} \rightarrow \Omega _{b'}^{(N')}\). \(\square \)
Going back to the proof of Proposition A.1, we see that Lemma A.2 permits to define a unique action of the quotient \(G = \frac{\Gamma }{\Gamma '}\) on \(X'\), compatible with its natural action on \(U'\). We then let \(X:= \frac{X'}{G}\). The following lemma ends the proof of Proposition A.1, and clarifies the link with the construction of [2].
Lemma A.3
The variety X defined above does not depend on the choice of \(\Gamma '\). When \(\Gamma \) is arithmetic, X coincides with the toroidal compactification of U constructed in [2].
Proof
Let \(\Gamma ', \Gamma '' \subset \Gamma \) be two neat lattices of finite index, and let us show that the varieties constructed from \(\Gamma '\) and \(\Gamma ''\) are the same. Since \(\Gamma \cap \Gamma '\) also has finite index in \(\Gamma \), we may assume \(\Gamma '' \subset \Gamma '\). By Lemma A.2, the action of two lattices \(\Gamma '' \subset \Gamma '\) are compatible with each other over each set \(\Omega _b^{(N)}\), which suffices to prove the first point.
Let us prove the second point. The construction of the toroidal compactification of [2] depends on a certain choice of \(\Gamma \)-admissible polyhedra for each rational boundary component (see [2, Definition 5.1]). In the case of the ball, since \(\dim _{{\mathbb {R}}} U_b = 1\) for any \(b \in \partial {\mathbb {B}}^n\), there is only one such possible choice (cf. [loc. cit., Theorem 4.1.(2)]). The claim now follows from the functoriality of compatible toroidal compactifications (see [28, Lemma 2.6]), since “choices” of polyhedra admissible for two lattices \(\Gamma ' \subset \Gamma \) are thus automatically compatible with each other. \(\square \)
Note that even though this construction of X is well adapted to our purposes, it should not be used to define X as an orbifold, as it has the drawback of producing artificial ramification orders along the boundary components of X. As explained in [21], a better way of proceeding is to construct directly open neighborhoods of the components of \(X - U\) as stacks, before glueing them to U.
1.2 A.2. Main results
Let us first begin with the following lemma.
Lemma A.4
Let Y be the toroidal compactification of the ball quotient \(U:=\frac{{\mathbb {B}}^n}{\Gamma }\) by a torsion free lattice \(\Gamma \subset PU(n,1)\) whose parabolic isometries are all unipotent. Let X be another projective compactification of U, and assume that \(X\) has at most klt singularities.
Then the identity map of U extends to a birational morphism \(f:X\rightarrow Y\).
Proof
The identity map of U defines a birational map \(f:X\dashrightarrow Y\). Assume by contradiction that f is not regular. One can take a resolution of indeterminacies \(\mu :{\tilde{X}}\rightarrow X\) for f so that \(\mu |_{\mu ^{-1}(U)}:\mu ^{-1}(U)\xrightarrow {\sim }U\) is an isomorphism:
By the rigidity result (see [17, Chapter 3, Lemma 1.15]), there is a fiber \(\mu ^{-1}(z)\) with \(z\in D\) which is not contracted by \({\tilde{f}}\). Clearly, we have \({\tilde{f}}(\mu ^{-1}(z)) \subset Y - U\).
Since X has klt singularities, the work of Hacon–McKernan [29] implies that every fiber of \(\mu \) is rationally chain connected. Thus, \({\widetilde{f}}(\mu ^{-1}(z))\) is a point since abelian varieties do not contain rational curves. This gives a contradiction. \(\square \)
Remark A.5
If we make the more restrictive hypothesis that \(X\) has at most quotient singularities, we can replace the use of [29] by the work of Kollar [36], which implies that each fiber of \(\mu \) is simply connected. As \(Y-U\) is a disjoint union of abelian varieties, this also implies that the image of \({\tilde{f}} : \mu ^{-1}(z) \rightarrow Y - U\) must be a point.
Let us introduce a natural class of pairs under which our rigidity theorem will hold.
Definition A.6
Let (X, D) be a pair consisting of normal algebraic variety and a reduced divisor. We say that (X, D) has algebraic quotient singularities if it admits a finite affine cover \((X_i)_{i \in I}\), such that each \((X_i, D \cap X_i)\) is the quotient of a smooth SNC pair \((U_i, D_i)\) by a finite group \(G_i\) leaving \(D_i\) invariant.
Note that for any lattice \(\Gamma \subset \mathrm{Aut}({\mathbb {B}}^n)\), if X is the toroidal compactification of \(U = \frac{{\mathbb {B}}^n}{\Gamma }\) described in Section 1, then \((X, X - U)\) has algebraic quotient singularities.
We can now state our main result as follows.
Theorem A.7
Let \(U:=\frac{{\mathbb {B}}^n}{\Gamma }\) be an n-dimensional ball quotient by a torsion free lattice \(\Gamma \subset PU(n,1)\). Let X be a klt compactification of U, and let \(D:=X-U\).
Let \(D^{(1)} \subset D\) be the divisorial part of D. If the Kähler–Einstein metric \(\omega \) on the bundle \(T_{X}(-\log D^{(1)})|_{U}\) is adapted to log order near the generic point of any component of \(D^{(1)}\), then (X, D) identifies with the toroidal compactification of U.
Remark A.8
-
(1)
Under the more restrictive assumption that (X, D) has algebraic quotient singularities, the use of Lemma A.4 in our proof below can be made without appealing to the difficult result of [29] (see Remark A.5).
-
(2)
As an easy consequence of Theorem A.7, we can remark that there is no klt compactification X of U such that \(X - U\) has codimension \(\ge 2\).
Corollary A.9
With the same assumptions as in Theorem A.7, if X is smooth and D has simple normal crossings, then D is in fact smooth, and each component is a smooth quotient of an abelian variety A by some finite group acting freely on A.
Let us prove Theorem A.7. Let \(\Gamma '\subset \Gamma \) be a subgroup of finite index so that all parabolic elements of \(\Gamma '\) are unipotent. Writing \(U':=\frac{{\mathbb {B}}^n}{\Gamma '}\), this gives a finite étale surjective morphism \(U'\rightarrow U\).
Let \(X'\) be the normalization of X in the function field of \(U'\): this is a normal projective variety \(X'\) compactifying \(U'\), with a compatible finite surjective morphism \(\mu :X'\rightarrow X\) (see e.g. [1, Chapter 12, §9]). Since klt singularities are preserved under finite surjective morphisms, the variety \(X'\) has at most klt singularities (see [31, Corollary 5.20]).
Remark A.10
If \((X, D)\) has algebraic quotient singularities, one sees easily that this is also the case for \(X'\). To see this, form the fiber product \(Z' = Z \times _{X} X'\), where \(Z \rightarrow X\) is an affine covering as in Definition A.6. By [37, Theorem 2.23], the variety \(Z'\), endowed with it natural boundary divisor, has algebraic quotient singularities. Finally, Lemma A.14 shows that \(Z' \rightarrow X'\) is a quotient map, which gives the result.
Let \(Y'\) be the toroidal compactification of \(U'\), so that the boundary \(A := Y' - U'\) is a smooth divisor.
Lemma A.11
The identity map on \(U'\) extends as an isomorphism \(f : X' \rightarrow Y'\). In particular, there is a finite surjective morphism \( g:Y'\rightarrow X, \) which identifies with the étale and surjective map \(U' \rightarrow U\) over \(X-D\).
Proof
Since \(X'\) is klt, Lemma A.4 shows that the identity map of \(U'\) extends to a birational morphism \(f:X'\rightarrow Y'\). Assume by contradiction that f is not an isomorphism. As \(Y'\) is smooth, it follows from [31, Corollary 2.63] that the exceptional set \(\mathrm{Ex}(f)\) is of pure codimension one. Thus, the birational morphism f must contract an irreducible divisorial component \(E\) of the boundary \(D':= X'-U'\).
Denote by \(D^{\mathrm{sing}}\) the singular locus of D, and let \(\omega ':=\mu ^*\omega \), be the canonical Kähler Einstein metric on \(U'\). Lemma A.12 below shows that \(\omega '\) is adapted to log-order for \(T_{X'^\circ }(-\log E^\circ )\), where \(X'^\circ :=\mu ^{-1}(X-D^{\mathrm{sing}})\), and \(E^\circ :=X'^\circ \cap E\). We are going to derive a contradiction with the fact the E is contracted. Let \(A_1\) be the component of A containing f(E). We can take admissible coordinates \(({\mathcal {W}};z_1,\ldots ,z_n)\) and \(({\mathcal {U}};w_1,\ldots ,w_n)\) centered at some well-chosen \(x'\in E\cap X'^\circ \) and \(y:=f(x')\in A_1\) respectively so that \(f({\mathcal {W}})\subset {\mathcal {U}}\), and \(f|_{E}:E\rightarrow f(E)\) is smooth at \(x'\). Denote by \( (f_1(z),\ldots ,f_n(z))\) the expression of f within these coordinates. Then if the admissible coordinates are chosen properly, one has
where \(g_1(z),\ldots ,g_k(z)\) are holomorphic functions defined on \({\mathcal {W}}\) so that \(g_i(z)\ne 0\) and \(m_{i}\ge 1\) for \(i=1,\ldots ,k\). Since E is exceptional, one has \(k\ge 2\). By the norm estimate in [44, eq. (8) on p. 338], the Kähler–Einstein metric \(\omega \) for \(T_{Y}(-\log A)|_{U}\) is adapted to log order. More precisely, one has
Since \( f^*d\log w_2=m_2d\log z_1+d\log g_2(w), \) one thus has the following norm estimate
for some constants \(C>0\). Since \(d\log z_1\) is a local nowhere vanishing section for \(\Omega _{X'}^1(\log D')\), we conclude that the metric \(\omega '^{-1}\) for \(\Omega _{X'^\circ }^{1}(-\log D'^\circ )\) is not adapted to log order, and so is \(\omega '\) for \(T_{X'^\circ }(-\log D'^\circ )\). This gives a contradiction, and ends the proof of the lemma. \(\square \)
Lemma A.12
With the notations of the proof of Lemma A.11, the metric \(\omega '\) is adapted to log-order for \(T_{X'^\circ }(-\log E^\circ )\).
Proof
Write \({\mathcal {W}}:=\mu ^{-1}({\mathcal {V}})\). Since \(\mu |_{{\mathcal {W}}-D'}:{\mathcal {W}}-D'\rightarrow {\mathcal {V}}-D\) is a finite unramified cover, the image of \((\mu |_{{\mathcal {W}}-D'})_*\big (\pi _1({\mathcal {W}}-D')\big )\) is a subgroup of \(\pi _1({\mathcal {W}}-D)\simeq {\mathbb {Z}}\) of index m. Letting \(\nu (z_{1}, \cdots , z_{n}) = (z_{1}^{m}, z_{2}, \ldots , z_{n})\), one has thus the following commutative diagram
so that \(h^\circ _{\Delta ^*\times \Delta ^{n-1}}:\Delta ^*\times \Delta ^{n-1}\rightarrow {\mathcal {W}}\cap U'\) is an isomorphism. By the Riemann removable singularities theorem, h extends to a holomorphic map \(h:\Delta ^n\rightarrow {\mathcal {W}}\), which is easily seen to be surjective with finite fibers. Hence h is moreover biholomorphic. \(({\mathcal {W}};z_1,\ldots ,z_n;h)\) is therefore an admissible coordinate centered at \(x'\) with \((z_1=0)={\mathcal {W}}\cap D'\) so we can now identify \(\mu \) with \(\nu \). Hence,
and the frame \((d\log z_1,dz_2,\ldots ,dz_n)\) for \(\Omega _{X'}^1(\log D')|_{{\mathcal {W}}}\) is adapted to log order. This shows that the metric \(\omega '\) is adapted to log order for \(T_{X'^\circ }(-\log D'^\circ )\). \(\square \)
We can now conclude the case discussed in Corollary A.9, where (X, D) is assumed to be a smooth log-pair. Since the boundary of \(Y' - U'\) is smooth, this implies that D must also be smooth. Moreover, for each connected component \(A_i\) of A, there is a connected component \(D_j\) of D so that \(g|_{A_i}:A_i\rightarrow D_j\) is a finite surjective morphism, which is also étale by the local description of \(\mu \) given in the proof of Lemma A.12. Hence in this case, \(D_i\) is a smooth quotient of an abelian variety by the free action of some finite group \(G_i\). This suffices to establish Corollary A.9.
The proof of Theorem A.7 will be complete with the following lemma.
Lemma A.13
The variety X identifies with the quotient of \(Y'\) by the natural action of \(G = \frac{\Gamma }{\Gamma '}\). In particular, \(X \cong Y\).
Proof
This result comes right away from Lemma A.14 below, taking \(M = Y'\), \(N = X\), and \(G = {\mathcal {G}}\). Remark that we have \(R(Y')^G = R(U')^G = R(U) = R(X)\) since \(U = \frac{U'}{G}\). For the second statement, remark that by Proposition A.1, the toroidal compactification Y of U also identifies with the quotient \(\frac{Y'}{G}\). Thus, there is an isomorphism \(Y \cong X\) compatible with the identity on U. Theorem A.7 is proved. \(\square \)
Lemma A.14
Let \(f : M \rightarrow N\) be a finite surjective morphism between two normal reduced schemes. Assume that M is acted upon by a finite groupoid \({\mathcal {G}}\), and that f is \({\mathcal {G}}\)-invariant. Suppose in addition that \(R(M)^{{\mathcal {G}}} = R(N)\), where R(M), R(N) are the rings of rational functions on M, N. Then N is the quotient of M by \({\mathcal {G}}\).
Proof
It suffices to show that \(f_*({\mathcal {O}}_M)^{{\mathcal {G}}} = {\mathcal {O}}_N\). This is a local statement on the base, so we may assume that \(N = \mathrm{Spec}\, A\), \(M = \mathrm{Spec}\, B\), with A is integrally closed. We have \(B^{{\mathcal {G}}} \subset R(B)^{{\mathcal {G}}} = R(A)\) by assumption. Since \(A \subset B\) is finite, and A is integrally closed, this implies \(B^{{\mathcal {G}}} \subset A\), as required. \(\square \)
Rights and permissions
About this article
Cite this article
Deng, Y., Cadorel, B. A characterization of complex quasi-projective manifolds uniformized by unit balls. Math. Ann. 384, 1833–1881 (2022). https://doi.org/10.1007/s00208-021-02334-z
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00208-021-02334-z