• Source: Hirzebruch surface
  • In mathematics, a Hirzebruch surface is a ruled surface over the projective line. They were studied by Friedrich Hirzebruch (1951).


    Definition


    The Hirzebruch surface




    Σ

    n




    {\displaystyle \Sigma _{n}}

    is the





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    -bundle (a projective bundle) over the projective line





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    , associated to the sheaf





    O





    O


    (

    n
    )
    .


    {\displaystyle {\mathcal {O}}\oplus {\mathcal {O}}(-n).}

    The notation here means:





    O


    (
    n
    )


    {\displaystyle {\mathcal {O}}(n)}

    is the n-th tensor power of the Serre twist sheaf





    O


    (
    1
    )


    {\displaystyle {\mathcal {O}}(1)}

    , the invertible sheaf or line bundle with associated Cartier divisor a single point. The surface




    Σ

    0




    {\displaystyle \Sigma _{0}}

    is isomorphic to





    P


    1


    ×


    P


    1




    {\displaystyle \mathbb {P} ^{1}\times \mathbb {P} ^{1}}

    ; and




    Σ

    1




    {\displaystyle \Sigma _{1}}

    is isomorphic to the projective plane





    P


    2




    {\displaystyle \mathbb {P} ^{2}}

    blown up at a point, so it is not minimal.


    = GIT quotient

    =
    One method for constructing the Hirzebruch surface is by using a GIT quotient: 21 :




    Σ

    n


    =
    (


    C


    2



    {
    0
    }
    )
    ×
    (


    C


    2



    {
    0
    }
    )

    /

    (


    C





    ×


    C





    )


    {\displaystyle \Sigma _{n}=(\mathbb {C} ^{2}-\{0\})\times (\mathbb {C} ^{2}-\{0\})/(\mathbb {C} ^{*}\times \mathbb {C} ^{*})}

    where the action of





    C





    ×


    C







    {\displaystyle \mathbb {C} ^{*}\times \mathbb {C} ^{*}}

    is given by



    (
    λ
    ,
    μ
    )

    (

    l

    0


    ,

    l

    1


    ,

    t

    0


    ,

    t

    1


    )
    =
    (
    λ

    l

    0


    ,
    λ

    l

    1


    ,
    μ

    t

    0


    ,

    λ


    n


    μ

    t

    1


    )

    .


    {\displaystyle (\lambda ,\mu )\cdot (l_{0},l_{1},t_{0},t_{1})=(\lambda l_{0},\lambda l_{1},\mu t_{0},\lambda ^{-n}\mu t_{1})\ .}

    This action can be interpreted as the action of



    λ


    {\displaystyle \lambda }

    on the first two factors comes from the action of





    C







    {\displaystyle \mathbb {C} ^{*}}

    on





    C


    2



    {
    0
    }


    {\displaystyle \mathbb {C} ^{2}-\{0\}}

    defining





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    , and the second action is a combination of the construction of a direct sum of line bundles on





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    and their projectivization. For the direct sum





    O





    O


    (

    n
    )


    {\displaystyle {\mathcal {O}}\oplus {\mathcal {O}}(-n)}

    this can be given by the quotient variety: 24 





    O





    O


    (

    n
    )
    =
    (


    C


    2



    {
    0
    }
    )
    ×


    C


    2



    /



    C







    {\displaystyle {\mathcal {O}}\oplus {\mathcal {O}}(-n)=(\mathbb {C} ^{2}-\{0\})\times \mathbb {C} ^{2}/\mathbb {C} ^{*}}

    where the action of





    C







    {\displaystyle \mathbb {C} ^{*}}

    is given by



    λ

    (

    l

    0


    ,

    l

    1


    ,

    t

    0


    ,

    t

    1


    )
    =
    (
    λ

    l

    0


    ,
    λ

    l

    1


    ,

    λ

    0



    t

    0


    =

    t

    0


    ,

    λ


    n



    t

    1


    )


    {\displaystyle \lambda \cdot (l_{0},l_{1},t_{0},t_{1})=(\lambda l_{0},\lambda l_{1},\lambda ^{0}t_{0}=t_{0},\lambda ^{-n}t_{1})}

    Then, the projectivization




    P

    (


    O





    O


    (

    n
    )
    )


    {\displaystyle \mathbb {P} ({\mathcal {O}}\oplus {\mathcal {O}}(-n))}

    is given by another





    C







    {\displaystyle \mathbb {C} ^{*}}

    -action: 22  sending an equivalence class



    [

    l

    0


    ,

    l

    1


    ,

    t

    0


    ,

    t

    1


    ]



    O





    O


    (

    n
    )


    {\displaystyle [l_{0},l_{1},t_{0},t_{1}]\in {\mathcal {O}}\oplus {\mathcal {O}}(-n)}

    to



    μ

    [

    l

    0


    ,

    l

    1


    ,

    t

    0


    ,

    t

    1


    ]
    =
    [

    l

    0


    ,

    l

    1


    ,
    μ

    t

    0


    ,
    μ

    t

    1


    ]


    {\displaystyle \mu \cdot [l_{0},l_{1},t_{0},t_{1}]=[l_{0},l_{1},\mu t_{0},\mu t_{1}]}

    Combining these two actions gives the original quotient up top.


    = Transition maps

    =
    One way to construct this





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    -bundle is by using transition functions. Since affine vector bundles are necessarily trivial, over the charts




    U

    0


    ,

    U

    1




    {\displaystyle U_{0},U_{1}}

    of





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    defined by




    x

    i



    0


    {\displaystyle x_{i}\neq 0}

    there is the local model of the bundle




    U

    i


    ×


    P


    1




    {\displaystyle U_{i}\times \mathbb {P} ^{1}}

    Then, the transition maps, induced from the transition maps of





    O





    O


    (

    n
    )


    {\displaystyle {\mathcal {O}}\oplus {\mathcal {O}}(-n)}

    give the map




    U

    0


    ×


    P


    1




    |



    U

    1






    U

    1


    ×


    P


    1




    |



    U

    0






    {\displaystyle U_{0}\times \mathbb {P} ^{1}|_{U_{1}}\to U_{1}\times \mathbb {P} ^{1}|_{U_{0}}}

    sending



    (

    X

    0


    ,
    [

    y

    0


    :

    y

    1


    ]
    )

    (

    X

    1


    ,
    [

    y

    0


    :

    x

    0


    n



    y

    1


    ]
    )


    {\displaystyle (X_{0},[y_{0}:y_{1}])\mapsto (X_{1},[y_{0}:x_{0}^{n}y_{1}])}

    where




    X

    i




    {\displaystyle X_{i}}

    is the affine coordinate function on




    U

    i




    {\displaystyle U_{i}}

    .


    Properties




    = Projective rank 2 bundles over P1

    =
    Note that by Grothendieck's theorem, for any rank 2 vector bundle



    E


    {\displaystyle E}

    on





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    there are numbers



    a
    ,
    b


    Z



    {\displaystyle a,b\in \mathbb {Z} }

    such that



    E



    O


    (
    a
    )



    O


    (
    b
    )
    .


    {\displaystyle E\cong {\mathcal {O}}(a)\oplus {\mathcal {O}}(b).}

    As taking the projective bundle is invariant under tensoring by a line bundle, the ruled surface associated to



    E
    =


    O


    (
    a
    )



    O


    (
    b
    )


    {\displaystyle E={\mathcal {O}}(a)\oplus {\mathcal {O}}(b)}

    is the Hirzebruch surface




    Σ

    b

    a




    {\displaystyle \Sigma _{b-a}}

    since this bundle can be tensored by





    O


    (

    a
    )


    {\displaystyle {\mathcal {O}}(-a)}

    .


    Isomorphisms of Hirzebruch surfaces


    In particular, the above observation gives an isomorphism between




    Σ

    n




    {\displaystyle \Sigma _{n}}

    and




    Σ


    n




    {\displaystyle \Sigma _{-n}}

    since there is the isomorphism vector bundles





    O


    (
    n
    )

    (


    O





    O


    (

    n
    )
    )



    O


    (
    n
    )



    O




    {\displaystyle {\mathcal {O}}(n)\otimes ({\mathcal {O}}\oplus {\mathcal {O}}(-n))\cong {\mathcal {O}}(n)\oplus {\mathcal {O}}}



    = Analysis of associated symmetric algebra

    =
    Recall that projective bundles can be constructed using Relative Proj, which is formed from the graded sheaf of algebras






    i
    =
    0






    Sym

    i



    (


    O





    O


    (

    n
    )
    )


    {\displaystyle \bigoplus _{i=0}^{\infty }\operatorname {Sym} ^{i}({\mathcal {O}}\oplus {\mathcal {O}}(-n))}

    The first few symmetric modules are special since there is a non-trivial anti-symmetric




    Alt

    2




    {\displaystyle \operatorname {Alt} ^{2}}

    -module





    O





    O


    (

    n
    )


    {\displaystyle {\mathcal {O}}\otimes {\mathcal {O}}(-n)}

    . These sheaves are summarized in the table








    Sym

    0



    (


    O





    O


    (

    n
    )
    )



    =


    O







    Sym

    1



    (


    O





    O


    (

    n
    )
    )



    =


    O





    O


    (

    n
    )





    Sym

    2



    (


    O





    O


    (

    n
    )
    )



    =


    O





    O


    (

    2
    n
    )






    {\displaystyle {\begin{aligned}\operatorname {Sym} ^{0}({\mathcal {O}}\oplus {\mathcal {O}}(-n))&={\mathcal {O}}\\\operatorname {Sym} ^{1}({\mathcal {O}}\oplus {\mathcal {O}}(-n))&={\mathcal {O}}\oplus {\mathcal {O}}(-n)\\\operatorname {Sym} ^{2}({\mathcal {O}}\oplus {\mathcal {O}}(-n))&={\mathcal {O}}\oplus {\mathcal {O}}(-2n)\end{aligned}}}

    For



    i
    >
    2


    {\displaystyle i>2}

    the symmetric sheaves are given by








    Sym

    k



    (


    O





    O


    (

    n
    )
    )



    =



    i
    =
    0


    k





    O




    (
    n

    i
    )





    O


    (

    i
    n
    )









    O





    O


    (

    n
    )





    O


    (

    k
    n
    )






    {\displaystyle {\begin{aligned}\operatorname {Sym} ^{k}({\mathcal {O}}\oplus {\mathcal {O}}(-n))&=\bigoplus _{i=0}^{k}{\mathcal {O}}^{\otimes (n-i)}\otimes {\mathcal {O}}(-in)\\&\cong {\mathcal {O}}\oplus {\mathcal {O}}(-n)\oplus \cdots \oplus {\mathcal {O}}(-kn)\end{aligned}}}



    = Intersection theory

    =
    Hirzebruch surfaces for n > 0 have a special rational curve C on them: The surface is the projective bundle of





    O


    (

    n
    )


    {\displaystyle {\mathcal {O}}(-n)}

    and the curve C is the zero section. This curve has self-intersection number −n, and is the only irreducible curve with negative self intersection number. The only irreducible curves with zero self intersection number are the fibers of the Hirzebruch surface (considered as a fiber bundle over





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    ). The Picard group is generated by the curve C and one of the fibers, and these generators have intersection matrix





    [



    0


    1




    1



    n



    ]


    ,


    {\displaystyle {\begin{bmatrix}0&1\\1&-n\end{bmatrix}},}

    so the bilinear form is two dimensional unimodular, and is even or odd depending on whether n is even or odd.
    The Hirzebruch surface Σn (n > 1) blown up at a point on the special curve C is isomorphic to Σn+1 blown up at a point not on the special curve.


    = Toric variety

    =
    The Hirzebruch surface




    Σ

    n




    {\displaystyle \Sigma _{n}}

    can be given an action of the complex torus



    T
    =


    C





    ×


    C







    {\displaystyle T=\mathbb {C} ^{*}\times \mathbb {C} ^{*}}

    , with one





    C







    {\displaystyle \mathbb {C} ^{*}}

    acting on the base





    P


    1




    {\displaystyle \mathbb {P} ^{1}}

    with two fixed axis points, and the other





    C







    {\displaystyle \mathbb {C} ^{*}}

    acting on the fibers of the vector bundle





    O





    O


    (

    n
    )


    {\textstyle {\mathcal {O}}\oplus {\mathcal {O}}(-n)}

    , specifically on the first line bundle component, and hence on the projective bundle. This produces an open orbit of T, making




    Σ

    n




    {\displaystyle \Sigma _{n}}

    a toric variety. Its associated fan partitions the standard lattice





    Z


    2




    {\displaystyle \mathbb {Z} ^{2}}

    into four cones (each corresponding to a coordinate chart), separated by the rays along the four vectors:



    (
    1
    ,
    0
    )
    ,
    (
    0
    ,
    1
    )
    ,
    (
    0
    ,

    1
    )
    ,
    (

    1
    ,
    n
    )
    .


    {\displaystyle (1,0),(0,1),(0,-1),(-1,n).}

    All the theory above generalizes to arbitrary toric varieties, including the construction of the variety as a quotient and by coordinate charts, as well as the explicit intersection theory.
    Any smooth toric surface except





    P


    2




    {\displaystyle \mathbb {P} ^{2}}

    can be constructed by repeatedly blowing up a Hirzebruch surface at T-fixed points.


    See also


    Projective bundle


    References



    Barth, Wolf P.; Hulek, Klaus; Peters, Chris A.M.; Van de Ven, Antonius (2004), Compact Complex Surfaces, Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge., vol. 4, Springer-Verlag, Berlin, ISBN 978-3-540-00832-3, MR 2030225
    Beauville, Arnaud (1996), Complex algebraic surfaces, London Mathematical Society Student Texts, vol. 34 (2nd ed.), Cambridge University Press, ISBN 978-0-521-49510-3, MR1406314
    Hirzebruch, Friedrich (1951), "Über eine Klasse von einfachzusammenhängenden komplexen Mannigfaltigkeiten", Mathematische Annalen, 124: 77–86, doi:10.1007/BF01343552, hdl:21.11116/0000-0004-3A56-B, ISSN 0025-5831, MR 0045384, S2CID 122844063


    External links


    Manifold Atlas
    https://www.mathematik.uni-kl.de/~gathmann/class/alggeom-2002/alggeom-2002-c10.pdf
    https://mathoverflow.net/q/122952

Kata Kunci Pencarian: