This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Fano 44-folds with nef tangent bundle in positive characteristic

Yuta Takahashi  and  Kiwamu Watanabe Department of Mathematics, Faculty of Science and Engineering, Chuo University. 1-13-27 Kasuga, Bunkyo-ku, Tokyo 112-8551, Japan [email protected] [email protected]
Abstract.

In characteristic 0, the Campana-Peternell conjecture claims that the only smooth Fano variety with nef tangent bundle should be homogeneous. In this paper, we study the positive characteristic version of the Campana-Peternell conjecture. In particular, we give an affirmative answer for Fano 44-folds with nef tangent bundle and Picard number greater than one.

2010 Mathematics Subject Classification:
14J35, 14J45, 14M17, 14E30.
The author is partially supported by JSPS KAKENHI Grant Number 21K03170 and the Sumitomo Foundation Grant Number 190170.

1. Introduction

How can one compare two given smooth projective varieties? Since any smooth variety XX has the tangent bundle TXT_{X}, we often use the tangent bundle TXT_{X} to compare smooth varieties. In particular, the positivity of the tangent bundle imposes strong restrictions on the geometry of varieties. For instance, in the celebrated paper [27], Mori solved the famous Hartshorne conjecture. The Hartshorne conjecture states that a smooth projective variety XX defined over an algebraically closed field is the projective space if TXT_{X} is ample. As a generalization of the Hartshorne conjecture, Campana and Peternell studied complex smooth projective varieties with nef tangent bundle [4]. In this direction, Demailly-Peternell-Schneider [8] proved that any complex smooth projective variety with nef tangent bundle is, up to an étale cover, a Fano fiber space over an Abelian variety. As a consequence, the study of complex smooth projective varieties with nef tangent bundle can be reduced to that of Fano varieties. Moreover Campana and Peternell [4] conjectured that any complex smooth Fano variety with nef tangent bundle is homogeneous. This conjecture holds for varieties of dimension at most five, but in general this is widely open. We refer the reader to [29].

In [18], the second author and Kanemitsu proved an analogue of the theorem by Demailly-Peternell-Schneider in positive characteristic; thus the next step is to study smooth Fano varieties with nef tangent bundle in positive characteristic. When the dimension is at most three, this problem was studied by the second author [36]. In the present paper, we give a classification of Fano 44-folds with nef tangent bundle and Picard number greater than one:

Theorem 1.1.

Let XX be a smooth Fano 44-fold defined over an algebraically closed field. If the tangent bundle TXT_{X} is nef and the Picard number of XX is greater than one, then XX is isomorphic to one of the following:

  1. (i)

    3×1{\mathbb{P}}^{3}\times{\mathbb{P}}^{1};

  2. (ii)

    Q3×1Q^{3}\times{\mathbb{P}}^{1};

  3. (iii)

    2×2{\mathbb{P}}^{2}\times{\mathbb{P}}^{2};

  4. (iv)

    2×1×1{\mathbb{P}}^{2}\times{\mathbb{P}}^{1}\times{\mathbb{P}}^{1};

  5. (v)

    (T2)×1{\mathbb{P}}(T_{{\mathbb{P}}^{2}})\times{\mathbb{P}}^{1}, where T2T_{{\mathbb{P}}^{2}} is the tangent bundle of 2{\mathbb{P}}^{2};

  6. (vi)

    1×1×1×1{\mathbb{P}}^{1}\times{\mathbb{P}}^{1}\times{\mathbb{P}}^{1}\times{\mathbb{P}}^{1};

  7. (vii)

    (𝒩){\mathbb{P}}({\mathcal{N}}) with a null-correlation bundle 𝒩{\mathcal{N}} on 3{\mathbb{P}}^{3} (see Definition 3.2).

In particular, XX is a homogeneous variety with reduced stabilizer.

In characteristic 0, this was proved by Camapana and Peternell [5]. However there are some difficulties to study this kind of classification problem in positive characteristic. For instance, the proof of [5] heavily depends on the Kodaira vanishing theorem and Hodge theory, which unfortunately fail in positive characteristic. We shall give a characteristic-free proof of [5].

The contents of this paper are organized as follows. In Section 2, we recall the background of our problem. We also review some known properties of Fano varieties with nef tangent bundle, paying special attention to some results in [18]. In Section 3, we shall study Fano varieties with nef tangent bundle which admit a projective bundle structure; this study plays a crucial role in the proof of Theorem 1.1. In Section 4, we will give a proof of Theorem 1.1.

2. Preliminaries

Notations

Let kk be an algebraically closed field of characteristic p0p\geq 0. Throughout this paper, we work over kk and use standard notations as in [10, 19, 21, 23, 24]. For a smooth projective variety XX, we also use the following notations:

  • We denote by TXT_{X} the tangent bundle of XX.

  • We denote by Ak(X)=Ank(X)A_{k}(X)=A^{n-k}(X) the group of rational equivalence classes of algebraic kk-cycles on XX. We denote by A(X):=kAk(X)A(X):=\bigoplus_{k}A_{k}(X) the Chow ring of XX.

  • We denote by N1(X)N_{1}(X) the group of numerical equivalence classes of algebraic 11-cycles with real coefficients on XX. The dimension dimN1(X)\dim_{{\mathbb{R}}}N_{1}(X) as an {\mathbb{R}}-vector space is called the Picard number of XX and we denote it by ρX\rho_{X}.

  • We say that a smooth projective variety XX is Fano if KX-K_{X} is ample. For a smooth Fano variety XX, the pseudoindex ιX\iota_{X} of XX is the minimal anticanonical degree of rational curves on XX.

  • An FF-bundle is a smooth morphism f:YXf:Y\to X between smooth projective varieties whose fibers are isomorphic to FF.

  • An elementary contraction means a contraction of an extremal ray.

  • For a vector bundle {\mathcal{E}} (resp. i{\mathcal{E}}_{i}) on XX, we denote the tautological divisor of (){\mathbb{P}}({\mathcal{E}}) (resp. (i){\mathbb{P}}({\mathcal{E}}_{i})) by ξ{\xi}_{{\mathcal{E}}} (resp. ξi{\xi}_{{\mathcal{E}}_{i}}). When no confusion is likely, we also simply denote the divisor ξ{\xi}_{{\mathcal{E}}} (resp. ξi{\xi}_{{\mathcal{E}}_{i}}) by ξ\xi (resp. ξi{\xi}_{i}).

  • For a rank two vector bundle {\mathcal{E}} on XX, we denote the ii-th Chern class of {\mathcal{E}} by ci()c_{i}({\mathcal{E}}). When A1(X)A^{1}(X) and A2(X)A^{2}(X) are isomorphic to {\mathbb{Z}}, there exist an effective divisor HH and an effective 22-cocycle LL on XX such that A1(X)[H]A^{1}(X)\cong{\mathbb{Z}}[H] and A2(X)[L]A^{2}(X)\cong{\mathbb{Z}}[L]; then we consider c1()c_{1}({\mathcal{E}}) and c2()c_{2}({\mathcal{E}}) as integers c1c_{1} and c2c_{2}, that is, c1()=c1HA1(X)c_{1}({\mathcal{E}})=c_{1}H\in A^{1}(X) and c2()=c2LA2(X)c_{2}({\mathcal{E}})=c_{2}L\in A^{2}(X). In this setting, we say that {\mathcal{E}} is normalized if c1=0c_{1}=0 or 1-1. We also say that {\mathcal{E}} is stable (resp. semistable) if for every invertible subsheaf {\mathcal{L}} of {\mathcal{E}}, c1()<12c1()c_{1}({\mathcal{L}})<\dfrac{1}{2}c_{1}({\mathcal{E}}) (resp. c1()12c1()c_{1}({\mathcal{L}})\leq\dfrac{1}{2}c_{1}({\mathcal{E}})).

  • For a vector bundle {\mathcal{E}} on XX, we say that {\mathcal{E}} is Fano if (){\mathbb{P}}({\mathcal{E}}) is a Fano variety.

  • For a vector bundle {\mathcal{E}} on XX, we say that {\mathcal{E}} is numerically flat if {\mathcal{E}} and its dual {\mathcal{E}}^{\vee} are nef (equivalently {\mathcal{E}} and det()\mathop{\rm det}\nolimits({\mathcal{E}}^{\vee}) are nef).

  • For a projective variety XX, we denote by RatCurvesn(X){\rm RatCurves}^{n}(X) the family of rational curves on XX (see [19, II Definition 2.11]).

  • We denote by n{\mathbb{P}}^{n} and QnQ^{n} projective nn-space and a smooth quadric hypersurface in n+1{\mathbb{P}}^{n+1} respectively.

2.1. Background of the Problem

Let XX be a smooth projective variety with nef tangent bundle. By the decomposition theorem [18, Theorem 1.7], XX admits a smooth contraction φ:XM\varphi:X\to M such that

  • any fiber of φ\varphi is a smooth Fano variety with nef tangent bundle;

  • the tangent bundle TMT_{M} is numerically flat.

This result suggests to study two special cases:

Question 2.1 ([18, Question 1.8], [4, Conjecture 11.1], [36, Question 1]).

Let XX be a smooth projective variety with nef tangent bundle.

  1. (i)

    If XX is a Fano variety, then is XX a homogeneous space with reduced stabilizer?

  2. (ii)

    If TXT_{X} is numerically flat, then is XX an étale quotient of an Abelian variety?

In characteristic zero, for special varieties, including Fano varieties whose dimension is at most five, affirmative answers to the first question are known (see [4, 5, 13, 15, 16, 17, 26, 28, 32, 34, 35, 37]), and an affirmative answer to the second question also follows from the Beauville-Bogomolov decomposition. On the other hand, very little is known in positive characteristic; we refer the reader to [14, 22, 25, 36]. Here we only recall the following:

Theorem 2.2 ([4, 36]).

Let XX be a smooth Fano nn-fold with nef tangent bundle. If nn is at most three, then XX is one of the following:

  1. (i)

    XX is the nn-dimensional projective space n{\mathbb{P}}^{n};

  2. (ii)

    XX is an nn-dimensional hyperquadric QnQ^{n} (n=2,3)(n=2,3);

  3. (iii)

    X=2×1X={\mathbb{P}}^{2}\times{\mathbb{P}}^{1};

  4. (iv)

    X=1×1×1X={\mathbb{P}}^{1}\times{\mathbb{P}}^{1}\times{\mathbb{P}}^{1};

  5. (v)

    X=(T2)X={\mathbb{P}}(T_{{\mathbb{P}}^{2}}).

To give a classification of complex Fano varieties with Picard number greater than one, it is quite common to study extremal contractions, but in positive characteristic, the existence of a contraction of an extremal ray is not known in general. The following result states that there exists a contraction of an extremal ray for Fano varieties with nef tangent bundle:

Theorem 2.3 (a special case of [18, Theorem 1.5]).

Let XX be a smooth Fano variety XX with nef tangent bundle. Let RNE(X)R\subset{\mathop{\rm NE}\nolimits}(X) be an extremal ray. Then the contraction f:XYf:X\to Y of the ray RR exists and the following hold:

  1. (i)

    ff is smooth;

  2. (ii)

    any fiber FF of ff is again a smooth Fano variety with nef tangent bundle;

  3. (iii)

    YY is also a smooth Fano variety with nef tangent bundle;

  4. (iv)

    ρX=ρY+1\rho_{X}=\rho_{Y}+1 and ρF=1\rho_{F}=1.

Let XX be a smooth projective variety. We say that XX is rationally chain connected (resp. rationally connected) if two general points on XX can be connected by a connected chain of rational curves (resp. by a single rational curve); it follows from [3], [20, Theorem 3.3] that smooth Fano varieties are rationally chain connected (see also [19, Chapter V. Theorem 2.13]). We say that XX is separably rationally connected if there exists a rational curve f:1Xf:{\mathbb{P}}^{1}\to X such that fTXf^{\ast}T_{X} is ample. In general, if XX is separably rationally connected, then it is rationally connected; by definition, a rationally connected variety is rationally chain connected; moreover these notions coincide in characteristic zero, whereas there exists a rationally connected variety which is not separably rationally connected in characteristic p>0p>0 (see for instance [19, V. Exercise 5.19]). For varieties with nef tangent bundle, these notions coincide:

Theorem 2.4 ([18, Theorem 1.3, Theorem 1.6]).

For a smooth projective variety XX with nef tangent bundle, the following are equivalent to each other:

  1. (i)

    XX is separably rationally connected;

  2. (ii)

    XX is rationally connected;

  3. (iii)

    XX is rationally chain connected;

  4. (iv)

    XX is a Fano variety.

Moreover, if XX satisfies the above equivalent conditions, the Kleiman-Mori cone NE(X){\mathop{\rm NE}\nolimits}(X) is simplicial.

We also have the following:

Theorem 2.5 (a special case of [18, Corollary 1.4]).

For a smooth Fano variety XX with nef TXT_{X}, the following hold:

  1. (i)

    XX is algebraically simply connected;

  2. (ii)

    H1(X,𝒪X)=0H^{1}(X,{\mathcal{O}}_{X})=0;

  3. (iii)

    every numerically flat vector bundle on XX is trivial.

2.2. Minimal birational sections

In this subsection, we recall minimal birational sections whose idea appeared in [38, 18]. Let XX be a smooth Fano variety with nef tangent bundle. Assume that f:XYf:X\to Y is an extremal contraction and dimY>0\dim Y>0. Since ff is a composition of contractions of extremal rays, Theorem 2.3 tells us that YY and any fiber of ff are smooth Fano varieties with nef tangent bundle. By Theorem 2.4, we see that any fiber FF of ff is separably rationally connected.

Definition 2.6.

Under the above notation, let CYC\subset Y be a rational curve. We call a rational curve C~X\tilde{C}\subset X a birational section of ff over CC if f|C~:C~Cf|_{\tilde{C}}:\tilde{C}\to C is birational. A birational section C~X\tilde{C}\subset X of ff over CC is minimal if the anticanonical degree KXC~-K_{X}\cdot\tilde{C} is minimal among birational sections of ff over CC.

Let us take a rational curve Y\ell\subset Y such that KY=ιY-K_{Y}\cdot\ell=\iota_{Y} and let 1Y{\mathbb{P}}^{1}\to\ell\subset Y be the normalization of \ell. We consider the fiber product:

X\textstyle{X_{\ell}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i\scriptstyle{i}f\scriptstyle{f_{\ell}}X\textstyle{X\ignorespaces\ignorespaces\ignorespaces\ignorespaces}f\scriptstyle{f}1\textstyle{{\mathbb{P}}^{1}\ignorespaces\ignorespaces\ignorespaces\ignorespaces}Y\textstyle{Y}

By the theorem of de Jong and Starr [7], ff_{\ell} admits a section ~\tilde{\ell}. Let us denote i(~)i(\tilde{\ell}) by ~X\tilde{\ell}_{X}; then f|~X:~Xf|_{\tilde{\ell}_{X}}:\tilde{\ell}_{X}\to\ell is birational; thus ~X\tilde{\ell}_{X} is a birational section of ff over \ell. This yields that there exists a minimal birational section of ff over \ell. As a consequence, we may find a rational curve 0Y\ell_{0}\subset Y and a minimal birational section of ff over 0\ell_{0} satisfying the following:

  • KY0=ιY-K_{Y}\cdot\ell_{0}=\iota_{Y};

  • KX0~X=min{deg(KX)~XKY=ιY,[]RatCurvesn(X)}-K_{X}\cdot\tilde{\ell_{0}}_{X}=\min\left\{\mathop{\rm deg}\nolimits_{(-K_{X})}\tilde{\ell}_{X}\mid-K_{Y}\cdot\ell=\iota_{Y},[\ell]\in{\rm RatCurves}^{n}(X)\right\}.

Let us consider a family of rational curves RatCurvesn(X){\mathcal{M}}\subset{\rm RatCurves}^{n}(X) containing [0~X][\tilde{\ell_{0}}_{X}]. By the same argument as in [38, Proposition 4.14], we see that {\mathcal{M}} is unsplit, that is, {\mathcal{M}} is proper as a scheme. Moreover [18, Proposition 4.4] implies that 0[0~X]{\mathbb{R}}_{\geq 0}[\tilde{\ell_{0}}_{X}] is an extremal ray and the contraction g:XZg:X\to Z of the ray 0[0~X]{\mathbb{R}}_{\geq 0}[\tilde{\ell_{0}}_{X}] is a smooth geometric quotient for {\mathcal{M}} in the sense of [2]. Summing up, we obtain the following:

Proposition 2.7.

Let XX be a smooth Fano variety with nef tangent bundle. Assume that f:XYf:X\to Y is an extremal contraction and dimY>0\dim Y>0. Then there exists an unsplit covering family of rational curves RatCurvesn(X){\mathcal{M}}\subset{\rm RatCurves}^{n}(X) such that

  • for any [~X][\tilde{\ell}_{X}]\in{\mathcal{M}}, f|~X:~Xf(~X)f|_{\tilde{\ell}_{X}}:\tilde{\ell}_{X}\to f(\tilde{\ell}_{X}) is birational and KYf(~X)=ιY-K_{Y}\cdot f(\tilde{\ell}_{X})=\iota_{Y};

  • 0[0~X]{\mathbb{R}}_{\geq 0}[\tilde{\ell_{0}}_{X}] is an extremal ray and the contraction g:XZg:X\to Z of the ray 0[0~X]{\mathbb{R}}_{\geq 0}[\tilde{\ell_{0}}_{X}] is a smooth geometric quotient for {\mathcal{M}}.

Definition 2.8 ([cf. [28, Definition 1]]).

Let XX be a smooth projective variety with nef tangent bundle. We say that XX is an FT-manifold if every elementary contraction of XX is a 1{\mathbb{P}}^{1}-bundle.

Example 2.9.

The variety (T2){\mathbb{P}}(T_{{\mathbb{P}}^{2}}) is isomorphic to a hyperplane section of a Segre 44-fold 2×28{\mathbb{P}}^{2}\times{\mathbb{P}}^{2}\subset{\mathbb{P}}^{8}. Since (T2){\mathbb{P}}(T_{{\mathbb{P}}^{2}}) admit two 1{\mathbb{P}}^{1}-bundle structures over 2{\mathbb{P}}^{2} and ρ(T2)=2\rho_{{\mathbb{P}}(T_{{\mathbb{P}}^{2}})}=2, it is an FT-manifold. The projective line 1{\mathbb{P}}^{1} is also a basic example of an FT-manifold.

Proposition 2.10.

Let XX be a smooth Fano variety with nef tangent bundle. Assume that f:XYf:X\to Y is an extremal contraction and dimY=1\dim Y=1. Then XX is isomorphic to a product of 1{\mathbb{P}}^{1} and a variety ZZ.

Proof.

We employ the notation as in Proposition 2.7. Remark that YY is 1{\mathbb{P}}^{1}. The contraction g:XZg:X\to Z is a 1{\mathbb{P}}^{1}-bundle; moreover any fiber of gg is a section of ff; this yields that XX is isomorphic to a product of 1{\mathbb{P}}^{1} and ZZ.  

Corollary 2.11.

Let MM be a smooth Fano variety with nef tangent bundle. Assume that f:MXf:M\to X is an extremal contraction onto an FT-manifold XX. Then MM is isomorphic to a product of XX and a variety YY.

Proof.

This follows from Proposition 2.10, Theorem 2.3 and the same argument as in [28, Proposition 5].  

2.3. Projective bundles

Definition 2.12 ([6, Definition 3.2]).

The (cohomological) Brauer group of a scheme YY is Br(Y):=He´t2(Y,𝔾m){\rm Br}(Y):=H_{\acute{e}t}^{2}(Y,{{\mathbb{G}}}_{m}).

Proposition 2.13.

Let f:XYf:X\to Y be a n{\mathbb{P}}^{n}-bundle. If the Brauer group Br(Y){\rm Br}(Y) vanishes, then there exists a vector bundle {\mathcal{E}} of rank n+1n+1 on YY such that X()X\cong{\mathbb{P}}({\mathcal{E}}).

Proof.

See for instance [12].  

Corollary 2.14.

Let f:XYf:X\to Y be a n{\mathbb{P}}^{n}-bundle. If YY is rational, then there exists a vector bundle {\mathcal{E}} of rank n+1n+1 on YY such that X()X\cong{\mathbb{P}}({\mathcal{E}}).

Proof.

By [6, Theorem 5.1.3, Proposition 5.2.2], we see that

Br(Y)Br(kn)Br(k).{\rm Br}(Y)\cong{\rm Br}({\mathbb{P}}^{n}_{k})\cong{\rm Br}(k).

Then [6, Corollary 1.2.4] implies that Br(k){\rm Br}(k) vanishes; thus our assertion follows from Proposition 2.13.  

3. Fano bundles over 2{\mathbb{P}}^{2}, 3{\mathbb{P}}^{3} and Q3Q^{3}

The major difficulty of the proof of Theorem 1.1 is to study the cases where a smooth Fano 44-fold with nef tangent bundle admits a 2{\mathbb{P}}^{2}-bundle structure over 2{\mathbb{P}}^{2} or a 1{\mathbb{P}}^{1}-bundle structure over 3{\mathbb{P}}^{3} and over Q3Q^{3}. In this section, we shall study such cases.

3.1. Rank three Fano bundles over 2{\mathbb{P}}^{2}

Proposition 3.1.

Let XX be a smooth Fano 44-fold with nef tangent bundle. Assume that f1:X2f_{1}:X\to{\mathbb{P}}^{2} is a 2{\mathbb{P}}^{2}-bundle. Then XX is isomorphic to 2×2{\mathbb{P}}^{2}\times{\mathbb{P}}^{2}.

Proof.

By Theorem 2.3, we may find another smooth elementary contraction f2:XX2f_{2}:X\to X_{2} besides f1f_{1}. Applying Theorem 2.2, Theorem 2.3 and [31], we see that f2:XX2f_{2}:X\to X_{2} is a 1{\mathbb{P}}^{1}-bundle over Q3Q^{3} or a 2{\mathbb{P}}^{2}-bundle over 2{\mathbb{P}}^{2}. We claim that f2:XX2f_{2}:X\to X_{2} is not a 1{\mathbb{P}}^{1}-bundle over Q3Q^{3}. To prove this, assume that f2:XX2f_{2}:X\to X_{2} is a 1{\mathbb{P}}^{1}-bundle over Q3Q^{3}. Then by Corollary 2.14, f1f_{1} and f2f_{2} are given by the projectivizations of vector bundles. Let us consider the Chow ring of XX. Since f1:X2f_{1}:X\to{\mathbb{P}}^{2} is a 2{\mathbb{P}}^{2}-bundle over 2{\mathbb{P}}^{2}, [9, Theorem 9.6] tells us that the rank of the A3(X)A^{3}(X) is three; however, since f2:XQ3f_{2}:X\to Q^{3} is a 1{\mathbb{P}}^{1}-bundle over Q3Q^{3}, [9, Theorem 9.6] tells us that the rank of the A3(X)A^{3}(X) is two; this is a contradiction. As a consequence, f2:XX2f_{2}:X\to X_{2} is a 2{\mathbb{P}}^{2}-bundle over 2{\mathbb{P}}^{2}. Applying [31], we conclude that XX is isomorphic to 2×2{\mathbb{P}}^{2}\times{\mathbb{P}}^{2}.  

3.2. Rank two Fano bundles over 3{\mathbb{P}}^{3}

Let us first recall the definition of the null-correlation bundle:

Definition 3.2 (see for instance [30, Section 4.2], [11, Example 8.4.1] and [39]).

Let {\mathcal{E}} be a rank 22 vector bundle on 3{\mathbb{P}}^{3}. We say that {\mathcal{E}} is a null-correlation bundle if it fits into an exact sequence

0𝒪3𝑠Ω3(2)(1)0,0\rightarrow{\mathcal{O}}_{{\mathbb{P}}^{3}}\xrightarrow{s}\Omega_{{\mathbb{P}}^{3}}(2)\rightarrow{\mathcal{E}}(1)\rightarrow 0,

where ss is a nowhere vanishing section of Ω3(2)\Omega_{{\mathbb{P}}^{3}}(2).

In this subsection, we prove the following:

Proposition 3.3.

Let XX be a smooth Fano 44-fold with nef tangent bundle. Assume that f1:X3f_{1}:X\to{\mathbb{P}}^{3} is a 1{\mathbb{P}}^{1}-bundle. Then XX is isomorphic to one of the following:

  1. (i)

    1×3{\mathbb{P}}^{1}\times{\mathbb{P}}^{3};

  2. (ii)

    (𝒩){\mathbb{P}}({\mathcal{N}}), where 𝒩{\mathcal{N}} is a null-correlation bundle.

Proof.

Let XX be a smooth Fano 44-fold with nef tangent bundle. Assume that f1:X3f_{1}:X\to{\mathbb{P}}^{3} is a 1{\mathbb{P}}^{1}-bundle. By Corollary 2.14, f1:X3f_{1}:X\to{\mathbb{P}}^{3} is given by the projectivization of a rank 22 vector bundle {\mathcal{E}} on 3{\mathbb{P}}^{3}, that is, f1:X=()3f_{1}:X={\mathbb{P}}({\mathcal{E}})\to{\mathbb{P}}^{3}. We assume that {\mathcal{E}} is normalized and consider its Chern classes c1()c_{1}({\mathcal{E}}) and c2()c_{2}({\mathcal{E}}) as integers c1c_{1} and c2c_{2} respectively. We denote by H1H_{1} the ample generator of Pic(3){\rm Pic}({\mathbb{P}}^{3}), by F1F_{1} a fiber of f1f_{1} and by ξ\xi the tautological divisor of (){\mathbb{P}}({\mathcal{E}}). By the same argument as in [33, Theorem 2.1] and [11, Example 8.4.1] (see also [39]), we see that one of the following holds:

  1. (i)

    {\mathcal{E}} is isomorphic to 𝒪3𝒪3{\mathcal{O}}_{{\mathbb{P}}^{3}}\oplus{\mathcal{O}}_{{\mathbb{P}}^{3}};

  2. (ii)

    {\mathcal{E}} is isomorphic to the null-correlation bundle 𝒩{\mathcal{N}};

  3. (iii)

    {\mathcal{E}} is isomorphic to 𝒪3𝒪3(1){\mathcal{O}}_{{\mathbb{P}}^{3}}\oplus{\mathcal{O}}_{{\mathbb{P}}^{3}}(-1);

  4. (iv)

    {\mathcal{E}} is isomorphic to 𝒪3(1)𝒪3(1){\mathcal{O}}_{{\mathbb{P}}^{3}}(-1)\oplus{\mathcal{O}}_{{\mathbb{P}}^{3}}(1);

  5. (v)

    {\mathcal{E}} is isomorphic to 𝒪3(2)𝒪3(1){\mathcal{O}}_{{\mathbb{P}}^{3}}(-2)\oplus{\mathcal{O}}_{{\mathbb{P}}^{3}}(1);

  6. (vi)

    {\mathcal{E}} is a stable bundle with (c1,c2)=(0,3)(c_{1},c_{2})=(0,3);

  7. (vii)

    {\mathcal{E}} is a stable bundle with (c1,c2)=(1,4)(c_{1},c_{2})=(-1,4).

If {\mathcal{E}} is isomorphic to 𝒪3𝒪3(1){\mathcal{O}}_{{\mathbb{P}}^{3}}\oplus{\mathcal{O}}_{{\mathbb{P}}^{3}}(-1), 𝒪3(1)𝒪3(1){\mathcal{O}}_{{\mathbb{P}}^{3}}(-1)\oplus{\mathcal{O}}_{{\mathbb{P}}^{3}}(1) or 𝒪3(2)𝒪3(1){\mathcal{O}}_{{\mathbb{P}}^{3}}(-2)\oplus{\mathcal{O}}_{{\mathbb{P}}^{3}}(1), then X=()X={\mathbb{P}}({\mathcal{E}}) admits a birational contraction, which contradicts to our assumption that the tangent bundle of XX is nef. To prove our assertion, it is enough to show that the cases (vi) and (vii) do not occur. In characteristic 0, it was proved in [33, Theorem 2.1], but we do not know whether their argument also holds in positive characteristic or not.

From now on, we prove that the cases (vi) and (vii) do not occur. Since XX is a smooth Fano variety with nef tangent bundle, Theorem 2.3 tells us that there exists another smooth elementary contraction f2:XX2f_{2}:X\to X_{2}. For any ample divisor H2Pic(X2)H_{2}\in{\rm Pic}(X_{2}), there exist integers a,ba,b such that

𝒪X(f2H2)𝒪X(aξ+bf1H1)Pic(X).{\mathcal{O}}_{X}(f_{2}^{\ast}H_{2})\cong{\mathcal{O}}_{X}(a\xi+bf_{1}^{\ast}H_{1})\in{\rm Pic}(X).

Let us first assume that {\mathcal{E}} is a rank 22 stable bundle on 3{\mathbb{P}}^{3} with (c1,c2)=(0,3)(c_{1},c_{2})=(0,3). We claim that aa and bb are not equal to 0. Taking the intersection number of aξ+bf1H1a\xi+bf_{1}^{\ast}H_{1} and F1F_{1}, we have

a=(aξ+bf1H1)F1=f2H2F1=H2f2(F1)>0.a=(a\xi+bf_{1}^{\ast}H_{1})\cdot F_{1}=f_{2}^{\ast}H_{2}\cdot F_{1}=H_{2}\cdot{f_{2}}_{\ast}(F_{1})>0.

Moreover, if b=0b=0, then we have 𝒪X(f2H2)𝒪X(aξ)Pic(X){\mathcal{O}}_{X}(f_{2}^{\ast}H_{2})\cong{\mathcal{O}}_{X}(a\xi)\in{\rm Pic}(X); in particular ξ\xi is nef; thus {\mathcal{E}} is nef and c1=0c_{1}=0. By Corollary 2.5 (iii), {\mathcal{E}} is isomorphic to 𝒪32{\mathcal{O}}_{{\mathbb{P}}^{3}}^{\oplus 2}; this is a contradiction. As a consequence, we see that b0b\neq 0.

Since we have ξ2+c2f1H12=0\xi^{2}+c_{2}f_{1}^{\ast}H_{1}^{2}=0, f1H14=0f_{1}^{\ast}H_{1}^{4}=0 and ξf1H13=1\xi\cdot f_{1}^{\ast}H_{1}^{3}=1, we also have

f1H14=0,ξf1H13=1,ξ2f1H12=0,ξ3f1H1=c2,ξ4=0.f_{1}^{\ast}H_{1}^{4}=0,\,\,\xi\cdot f_{1}^{\ast}H_{1}^{3}=1,\,\,\xi^{2}\cdot f_{1}^{\ast}H_{1}^{2}=0,\,\,\xi^{3}\cdot f_{1}^{\ast}H_{1}=-c_{2},\,\,\xi^{4}=0.

By using these equations, we have

0=(f2H2)4=(aξ+bf1H1)4=4ab(3a2+b2).0=(f_{2}^{\ast}H_{2})^{4}=(a\xi+bf_{1}^{\ast}H_{1})^{4}=4ab(-3a^{2}+b^{2}).

Since a,b0a,b\neq 0, we have

3a2+b2=0.-3a^{2}+b^{2}=0.

However this is a contradiction.

Secondly assume that {\mathcal{E}} is a rank 22 stable bundle on 3{\mathbb{P}}^{3} with (c1,c2)=(1,4)(c_{1},c_{2})=(-1,4). By the same argument as in the case where {\mathcal{E}} is stable and (c1,c2)=(0,3)(c_{1},c_{2})=(0,3), we have

a>0,f1H14=0,ξf1H13=1,ξ2f1H12=1,ξ3f1H1=3,ξ4=7.a>0,\,\,f_{1}^{\ast}H_{1}^{4}=0,\,\,\xi\cdot f_{1}^{\ast}H_{1}^{3}=1,\,\,\xi^{2}\cdot f_{1}^{\ast}H_{1}^{2}=-1,\,\,\xi^{3}\cdot f_{1}^{\ast}H_{1}=-3,\,\,\xi^{4}=7.

Thus we have an equation

0=(aξ+bf1H1)4=7a412a3b6a2b2+4ab3.0=(a\xi+bf_{1}^{\ast}H_{1})^{4}=7a^{4}-12a^{3}b-6a^{2}b^{2}+4ab^{3}.

Since aa is positive, we have

4(ba)36(ba)212(ba)+7=0.4\left(\dfrac{\,b\,}{a}\right)^{3}-6\left(\dfrac{\,b\,}{a}\right)^{2}-12\left(\dfrac{\,b\,}{a}\right)+7=0.

Then it follows from the rational root theorem that ba= 12\dfrac{\,b\,}{a}=\dfrac{\,1\,}{2}; this implies that 2ξ+f1H12\xi+f_{1}^{\ast}H_{1} is nef. Since X=()X={\mathbb{P}}({\mathcal{E}}) is a Fano variety, we have

0(2ξ+f1H1)(KX)3=(2ξ+f1H1)(2ξ+5f1H1)3=232.0\leq\left(2\xi+f_{1}^{\ast}H_{1}\right)\cdot\left(-K_{X}\right)^{3}=\left(2\xi+f_{1}^{\ast}H_{1}\right)\cdot\left(2\xi+5f_{1}^{\ast}H_{1}\right)^{3}=-232.

This is a contradiction. As a consequence, our assertion holds.  

3.3. Rank two Fano bundles over Q3Q^{3}

In this subsection, we prove the following:

Proposition 3.4.

Let XX be a smooth Fano 44-fold with nef tangent bundle. Assume that f1:XQ3f_{1}:X\to Q^{3} is a 1{\mathbb{P}}^{1}-bundle. Then XX does not admit another 1{\mathbb{P}}^{1}-bundle structure on Q3Q^{3}.

Proof.

Let XX be a smooth Fano 44-fold with nef tangent bundle. Assume that f1:XX1:=Q3f_{1}:X\to X_{1}:=Q^{3} is a 1{\mathbb{P}}^{1}-bundle. To prove Proposition 3.4, assume the contrary; then we have another 1{\mathbb{P}}^{1}-bundle structure f2:XX2:=Q3f_{2}:X\to X_{2}:=Q^{3} besides f1f_{1}. By Corollary 2.14, fi:XXif_{i}:X\to X_{i} (i=1,2i=1,2) is given by the projectivization of a rank 22 vector bundle i{\mathcal{E}}_{i} on Q3Q^{3}. Denoting by ξi\xi_{i} the divisor corresponding to the tautological line bundle 𝒪(i)(1){\mathcal{O}}_{{\mathbb{P}}({\mathcal{E}}_{i})}(1) and by HiH_{i} the ample generator of Pic(Xi){\rm Pic}(X_{i}), we have

Pic(X)[ξi][fiHi]for i=1, 2.{\rm Pic}(X)\cong{\mathbb{Z}}[\xi_{i}]\oplus{\mathbb{Z}}[f_{i}^{\ast}H_{i}]\,\,\,\,~{}\mbox{for~{}}i=1,\,2.

Remark that the Chow groups A1(Q3)A^{1}(Q^{3}) and A2(Q3)A^{2}(Q^{3}) are isomorphic to {\mathbb{Z}}; then we consider the Chern classes c1(1),c1(2),c2(1)c_{1}({\mathcal{E}}_{1}),c_{1}({\mathcal{E}}_{2}),c_{2}({\mathcal{E}}_{1}) and c2(2)c_{2}({\mathcal{E}}_{2}) as integers c1,c1,c2c_{1},c_{1}^{\prime},c_{2} and c2c_{2}^{\prime} respectively:

c1(1)=c1H1,c1(2)=c1H2,c2(1)=c22H12,c2(2)=c22H22.c_{1}({\mathcal{E}}_{1})=c_{1}H_{1},\,\,c_{1}({\mathcal{E}}_{2})=c_{1}^{\prime}H_{2},\,\,c_{2}({\mathcal{E}}_{1})=\dfrac{c_{2}}{2}H_{1}^{2},\,\,c_{2}({\mathcal{E}}_{2})=\dfrac{c_{2}^{\prime}}{2}H_{2}^{2}.

We may assume that i{\mathcal{E}}_{i}’s are normalized, that is, c1,c1{1,0}c_{1},c_{1}^{\prime}\in\{-1,0\}. We denote by FiF_{i} a fiber of fif_{i}.

By Proposition 2.7, f2f_{2} is nothing but the elementary contraction which contracts minimal birational sections of f1f_{1} over lines on Q3Q^{3}; this implies that

f1H1F2=1.f_{1}^{\ast}H_{1}\cdot F_{2}=1.

Then we obtain

2=KXF2=2ξ1F2+(3c1).2=-K_{X}\cdot F_{2}=2\xi_{1}\cdot F_{2}+(3-c_{1}).

Comparing the parity of both sides of this equation, we see that c1=1c_{1}=-1; thus we obtain

ξ1F2=1.\xi_{1}\cdot F_{2}=-1.

Since we have ξ12+ξ1f1H1+12c2f1H12=0\xi_{1}^{2}+\xi_{1}\cdot f_{1}^{\ast}H_{1}+\dfrac{1}{2}c_{2}f_{1}^{\ast}H_{1}^{2}=0, f1H14=0f_{1}^{\ast}H_{1}^{4}=0 and ξ1f1H13=2\xi_{1}\cdot f_{1}^{\ast}H_{1}^{3}=2, we also have

f1H14=0,ξ1f1H13=2,ξ12f1H12=2,ξ13f1H1=2c2,ξ14=2c22.f_{1}^{\ast}H_{1}^{4}=0,\,\,\xi_{1}\cdot f_{1}^{\ast}H_{1}^{3}=2,\,\,\xi_{1}^{2}\cdot f_{1}^{\ast}H_{1}^{2}=-2,\,\,\xi_{1}^{3}\cdot f_{1}^{\ast}H_{1}=2-c_{2},\,\,\xi_{1}^{4}=2c_{2}-2.

For the ample generator H2H_{2} of Pic(X2){\rm Pic}(X_{2}), there exist integers a,ba,b such that 𝒪X(f2H2)𝒪X(aξ1+bf1H1)Pic(X){\mathcal{O}}_{X}(f_{2}^{\ast}H_{2})\cong{\mathcal{O}}_{X}(a\xi_{1}+bf_{1}^{\ast}H_{1})\in{\rm Pic}(X). By this definition, (a,b)(0,0)(a,b)\neq(0,0). Moreover we have

0=(aξ1+bf1H1)F2=a+b.0=(a\xi_{1}+bf_{1}^{\ast}H_{1})\cdot F_{2}=-a+b.

Hence, we obtain a=b0a=b\neq 0; then we obtain

0=(ξ1+f1H1)4=2c2+2.0=(\xi_{1}+f_{1}^{\ast}H_{1})^{4}=-2c_{2}+2.

Hence we obtain c2=1c_{2}=1. Let us take integers α,β,γ,δ\alpha,\beta,\gamma,\delta as follows:

(1) {ξ1=αξ2+βf2H2f1H1=γξ2+δf2H2\begin{cases}\xi_{1}&=\alpha\xi_{2}+\beta f_{2}^{\ast}H_{2}\\ f_{1}^{\ast}H_{1}&=\gamma\xi_{2}+\delta f_{2}^{\ast}H_{2}\end{cases}

Remark that |αδβγ|=1|\alpha\delta-\beta\gamma|=1, because {ξi,fiHi}\{\xi_{i},f_{i}^{\ast}H_{i}\} is a {\mathbb{Z}}-basis of Pic(X){\rm Pic}(X) for i=1,2i=1,2. Since we have

1=ξ1F2=α,1=ξ1F1=ξ2F1+βf2H2F1-1=\xi_{1}\cdot F_{2}=\alpha,\quad 1=\xi_{1}\cdot F_{1}=-\xi_{2}\cdot F_{1}+\beta f_{2}^{\ast}H_{2}\cdot F_{1}

and

1=f1H1F2=γ,0=f1H1F1=ξ2F1+δf2H2F1,1=f_{1}^{\ast}H_{1}\cdot F_{2}=\gamma,\quad 0=f_{1}^{\ast}H_{1}\cdot F_{1}=\xi_{2}\cdot F_{1}+\delta f_{2}^{\ast}H_{2}\cdot F_{1},

we obtain

(2) {ξ1=ξ2+1+ξ2F1f2H2F1f2H2f1H1=ξ2ξ2F1f2H2F1f2H2\begin{cases}\xi_{1}&=-\xi_{2}+\dfrac{1+\xi_{2}\cdot F_{1}}{f_{2}^{\ast}H_{2}\cdot F_{1}}f_{2}^{\ast}H_{2}\\ f_{1}^{\ast}H_{1}&=\xi_{2}-\dfrac{\xi_{2}\cdot F_{1}}{f_{2}^{\ast}H_{2}\cdot F_{1}}f_{2}^{\ast}H_{2}\end{cases}

Then the equality |αδβγ|=1|\alpha\delta-\beta\gamma|=1 implies that f2H2F1=1f_{2}^{\ast}H_{2}\cdot F_{1}=1. Computing KXF1-K_{X}\cdot F_{1}, we see that c1=1c_{1}^{\prime}=-1 and ξ2F1=1\xi_{2}\cdot F_{1}=-1. Thus the equation (2) can be written as follows:

(3) {ξ1=ξ2f1H1=ξ2+f2H2\begin{cases}\xi_{1}&=-\xi_{2}\\ f_{1}^{\ast}H_{1}&=\xi_{2}+f_{2}^{\ast}H_{2}\end{cases}

Now we have

2=H23\displaystyle 2=H_{2}^{3} =\displaystyle= f1H1f2H23\displaystyle f_{1}^{\ast}H_{1}\cdot f_{2}^{\ast}H_{2}^{3}
=\displaystyle= f1H1(f1H1+ξ1)3\displaystyle f_{1}^{\ast}H_{1}\cdot\left(f_{1}^{\ast}H_{1}+\xi_{1}\right)^{3}
=\displaystyle= f1H14+3f1H13ξ1+3f1H12ξ12+f1H1ξ13=1\displaystyle f_{1}^{\ast}H_{1}^{4}+3f_{1}^{\ast}H_{1}^{3}\xi_{1}+3f_{1}^{\ast}H_{1}^{2}\xi_{1}^{2}+f_{1}^{\ast}H_{1}\xi_{1}^{3}=1

This is a contradiction.  

4. Proof of the main theorem

We prove Theorem 1.1. Let XX be a smooth Fano 44-fold defined over an algebraically closed field kk. Assume that the tangent bundle TXT_{X} is nef and the Picard number of XX is greater than one. By Theorem 2.3, we may find a smooth elementary contraction f:XYf:X\to Y. We denote any fiber of ff by FF; then by Theorem 2.3 again, FF and YY are smooth Fano varieties with nef tangent bundle, and we also have ρX=ρF+ρY\rho_{X}=\rho_{F}+\rho_{Y}, dimF+dimY=4\dim F+\dim Y=4 and dimF,dimY>0\dim F,\dim Y>0. If XX admits a contraction onto an FT-manifold WW, then it follows from Corollary 2.11 that XX is isomorphic to Z×WZ\times W for some variety ZZ; then ZZ is one of varieties in Theorem 2.2. Thus our assertion holds in this case. By Theorem 2.2, we may assume that YY is isomorphic to 2,3{\mathbb{P}}^{2},{\mathbb{P}}^{3} or Q3Q^{3}. Then by Corollary 2.11, Proposition 3.1, Proposition 3.3 and Proposition 3.4, we may conclude our assertion.

References

  • [1] Indranil Biswas and João Pedro Dos Santos. Triviality criteria for bundles over rationally connected varieties. J. Ramanujan Math. Soc., 28(4):423–442, 2013.
  • [2] Laurent Bonavero, Cinzia Casagrande, and Stéphane Druel. On covering and quasi-unsplit families of curves. J. Eur. Math. Soc. (JEMS), 9(1):45–57, 2007.
  • [3] Frédéric Campana. Connexité rationnelle des variétés de Fano. Ann. Sci. École Norm. Sup. (4), 25(5):539–545, 1992.
  • [4] Frédéric Campana and Thomas Peternell. Projective manifolds whose tangent bundles are numerically effective. Math. Ann., 289(1):169–187, 1991.
  • [5] Frédéric Campana and Thomas Peternell. 44-folds with numerically effective tangent bundles and second Betti numbers greater than one. Manuscripta Math., 79(3-4):225–238, 1993.
  • [6] Jean-Louis Colliot-Thélène and Alexei N. Skorobogatov. The Brauer-Grothendieck group, volume 71 of Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics [Results in Mathematics and Related Areas. 3rd Series. A Series of Modern Surveys in Mathematics]. Springer, Cham, [2021] ©2021.
  • [7] A. J. de Jong and J. Starr. Every rationally connected variety over the function field of a curve has a rational point. Amer. J. Math., 125(3):567–580, 2003.
  • [8] Jean-Pierre Demailly, Thomas Peternell, and Michael Schneider. Compact complex manifolds with numerically effective tangent bundles. J. Algebraic Geom., 3(2):295–345, 1994.
  • [9] David Eisenbud and Joe Harris. 3264 and all that—a second course in algebraic geometry. Cambridge University Press, Cambridge, 2016.
  • [10] Robin Hartshorne. Algebraic geometry. Springer-Verlag, New York-Heidelberg, 1977. Graduate Texts in Mathematics, No. 52.
  • [11] Robin Hartshorne. Stable vector bundles of rank 22 on 𝐏3{\bf P}^{3}. Math. Ann., 238(3):229–280, 1978.
  • [12] Takumi Murayama (https://math.stackexchange.com/users/116766/takumi murayama). Geometric fibers n{\mathbb{P}}^{n} + vanishing of brauer group implies projective bundle. Mathematics Stack Exchange. URL:https://math.stackexchange.com/q/2225438 (version: 2017-04-10).
  • [13] Jun-Muk Hwang. Rigidity of rational homogeneous spaces. In International Congress of Mathematicians. Vol. II, pages 613–626. Eur. Math. Soc., Zürich, 2006.
  • [14] Kirti Joshi. On varieties with trivial tangent bundle in characteristic p>0p>0. Nagoya Math. J., 242:35–51, 2021.
  • [15] Akihiro Kanemitsu. Fano nn-folds with nef tangent bundle and Picard number greater than n5n-5. Math. Z., 284(1-2):195–208, 2016.
  • [16] Akihiro Kanemitsu. Fano 5-folds with nef tangent bundles. Math. Res. Lett., 24(5):1453–1475, 2017.
  • [17] Akihiro Kanemitsu. Extremal rays and nefness of tangent bundles. Michigan Math. J., 68(2):301–322, 2019.
  • [18] Akihiro Kanemitsu and Kiwamu Watanabe. Projective varieties with nef tangent bundle in positive characteristic. Preprint arXiv:2012.09419, 2020.
  • [19] János Kollár. Rational curves on algebraic varieties, volume 32 of Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics [Results in Mathematics and Related Areas. 3rd Series. A Series of Modern Surveys in Mathematics]. Springer-Verlag, Berlin, 1996.
  • [20] János Kollár, Yoichi Miyaoka, and Shigefumi Mori. Rational connectedness and boundedness of Fano manifolds. J. Differential Geom., 36(3):765–779, 1992.
  • [21] János Kollár and Shigefumi Mori. Birational geometry of algebraic varieties, volume 134 of Cambridge Tracts in Mathematics. Cambridge University Press, Cambridge, 1998. With the collaboration of C. H. Clemens and A. Corti, Translated from the 1998 Japanese original.
  • [22] Adrian Langer. Generic positivity and foliations in positive characteristic. Adv. Math., 277:1–23, 2015.
  • [23] Robert Lazarsfeld. Positivity in algebraic geometry. I, volume 48 of Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics [Results in Mathematics and Related Areas. 3rd Series. A Series of Modern Surveys in Mathematics]. Springer-Verlag, Berlin, 2004. Classical setting: line bundles and linear series.
  • [24] Robert Lazarsfeld. Positivity in algebraic geometry. II, volume 49 of Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics [Results in Mathematics and Related Areas. 3rd Series. A Series of Modern Surveys in Mathematics]. Springer-Verlag, Berlin, 2004. Positivity for vector bundles, and multiplier ideals.
  • [25] V. B. Mehta and V. Srinivas. Varieties in positive characteristic with trivial tangent bundle. Compositio Math., 64(2):191–212, 1987. With an appendix by Srinivas and M. V. Nori.
  • [26] Ngaiming Mok. On Fano manifolds with nef tangent bundles admitting 1-dimensional varieties of minimal rational tangents. Trans. Amer. Math. Soc., 354(7):2639–2658 (electronic), 2002.
  • [27] Shigefumi Mori. Projective manifolds with ample tangent bundles. Ann. of Math. (2), 110(3):593–606, 1979.
  • [28] Roberto Muñoz, Gianluca Occhetta, Luis E. Solá Conde, and Kiwamu Watanabe. Rational curves, Dynkin diagrams and Fano manifolds with nef tangent bundle. Math. Ann., 361(3-4):583–609, 2015.
  • [29] Roberto Muñoz, Gianluca Occhetta, Luis E. Solá Conde, Kiwamu Watanabe, and Jarosław A. Wiśniewski. A survey on the Campana-Peternell conjecture. Rend. Istit. Mat. Univ. Trieste, 47:127–185, 2015.
  • [30] Christian Okonek, Michael Schneider, and Heinz Spindler. Vector bundles on complex projective spaces. Progress in Mathematics, 3. Birkhäuser, Boston, Mass., 1980.
  • [31] Ei-ichi Sato. Varieties which have two projective space bundle structures. J. Math. Kyoto Univ., 25(3):445–457, 1985.
  • [32] Luis E. Solá Conde and Jarosław A. Wiśniewski. On manifolds whose tangent bundle is big and 1-ample. Proc. London Math. Soc. (3), 89(2):273–290, 2004.
  • [33] Michał Szurek and Jarosław A. Wiśniewski. Fano bundles over 3{\mathbb{P}}^{3} and 3{\mathbb{Q}}_{3}. Pacific J. Math., 141(1):197–208, 1990.
  • [34] Kiwamu Watanabe. Fano 5-folds with nef tangent bundles and Picard numbers greater than one. Math. Z., 276(1-2):39–49, 2014.
  • [35] Kiwamu Watanabe. Fano manifolds with nef tangent bundle and large Picard number. Proc. Japan Acad. Ser. A Math. Sci., 91(6):89–94, 2015.
  • [36] Kiwamu Watanabe. Low-dimensional projective manifolds with nef tangent bundle in positive characteristic. Comm. Algebra, 45(9):3768–3777, 2017.
  • [37] Kiwamu Watanabe. Fano manifolds of coindex three admitting nef tangent bundle. Geom. Dedicata, 210:165–178, 2021.
  • [38] Kiwamu Watanabe. Positivity of the second exterior power of the tangent bundles. Adv. Math., 385:Paper No. 107757, 27, 2021.
  • [39] G. Pete Wever. The moduli of a class of rank 22 vector bundles on P3P^{3}. Nagoya Math. J., 84:9–30, 1981.