Documentation

DerivedAlgGeo.AlgebraicGeometry.Numerical.Examples.Surface.SmoothQuadricCharge

The divisorial central charge on a smooth quadric surface #

Over an algebraically closed field, the smooth quadric surface is Q = P^1 x P^1. Its real numerical divisor space has basis given by the two rulings f₁, f₂, with

f₁^2 = f₂^2 = 0, f₁.f₂ = 1.

This is the first example in which a scalar field B = beta H is not the intrinsic parameter. We model

H = h₁ f₁ + h₂ f₂, B = b₁ f₁ + b₂ f₂

independently and recover the explicit rank-two formula. The ample cone is recorded separately as h₁ > 0 and h₂ > 0, while the arithmetic central charge itself remains defined for every pair of real classes.

The numerical carrier uses coordinates (r, c, d, v) for ch(E) = (r, c f₁ + d f₂, v). Its real Chern character is induced from the rational graded ring through Surface.NumericalRealization, so this charge is a child of the general divisorial construction rather than an independent coordinate polynomial.

The arbitrary real scalar-extension presentation #

A real divisor class in the rational ruling basis, regarded as an element of ℝ ⊗[ℚ] A¹(Q)_ℚ.

Equations
  • One or more equations did not get rendered due to their size.
Instances For
    @[simp]

    The scalar-extended ruling coordinates realize to the corresponding arbitrary real class in N¹(Q)_ℝ.

    The scalar-extension realization reaches every real divisor class on the smooth quadric.

    Independent real B and omega in the scalar extension of the quadric's rational numerical ring.

    Equations
    • One or more equations did not get rendered due to their size.
    Instances For
      @[simp]

      Realizing scalar-extended parameters recovers the intrinsic ruling coordinates.

      The smooth-quadric child of the intrinsic divisorial central charge.

      Equations
      • One or more equations did not get rendered due to their size.
      Instances For

        The coordinate charge is exactly the canonical scalar-extension charge for every real pair B = (b₁,b₂) and omega = (h₁,h₂). Unlike the rational compatibility theorem below, this statement loses no real divisor classes.

        Bundle the geometric ampleness side condition without changing the charge arithmetic.

        Equations
        Instances For
          theorem AlgebraicGeometry.Numerical.Examples.SmoothQuadric.omega_square (h₁ h₂ : ℝ) :
          divisorSpace.pair (h₁, h₂) (h₁, h₂) = 2 * h₁ * h₂

          The polarization square is 2 h₁ h₂.

          theorem AlgebraicGeometry.Numerical.Examples.SmoothQuadric.omega_square_pos {h₁ h₂ : ℝ} (hh₁ : 0 < h₁) (hh₂ : 0 < h₂) :
          0 < divisorSpace.pair (h₁, h₂) (h₁, h₂)

          An ample class on the smooth quadric has positive square. This is a derived fact in the example, not the definition of ampleness in the core API.

          theorem AlgebraicGeometry.Numerical.Examples.SmoothQuadric.centralCharge_apply (h₁ h₂ b₁ b₂ : ℝ) (r c d v : ℤ) :
          (charge h₁ h₂ b₁ b₂) (r, c, d, v) = ↑(-↑v + b₁ * ↑d + b₂ * ↑c + (h₁ * h₂ - b₁ * b₂) * ↑r) + Complex.I * ↑(h₁ * ↑d + h₂ * ↑c - (h₁ * b₂ + h₂ * b₁) * ↑r)

          Explicit central charge for a smooth quadric in the ruling basis.

          For ch(E) = (r, c f₁ + d f₂, v), the real and imaginary parts are

          -v + b₁d + b₂c + (h₁h₂-b₁b₂)r, h₁d + h₂c - (h₁b₂+h₂b₁)r.

          The intrinsic real charge agrees with the generic numerical-ring charge for arbitrary rational ample polarization and arbitrary rational B-field.

          The direction f₁-f₂, orthogonal to the Segre polarization.

          Equations
          Instances For

            The inherited orthogonal wall slice #

            Linear parameterization of the direction transverse to the Segre class.

            Equations
            • One or more equations did not get rendered due to their size.
            Instances For

              The full rank-two wall slice of the quadric, as a child of the basis-free orthogonal-slice construction.

              Equations
              • One or more equations did not get rendered due to their size.
              Instances For

                The certificates of the slice #

                Walls/Divisorial/Slice.lean attaches IsHodge and IsGeometric to an orthogonal slice, and wallSlice carried neither. Both hold here, and the pieces were already in the file: the Segre class has positive square, the anti-diagonal direction has negative square, and ampleCone is exactly the set IsGeometric wants.

                On a hyperbolic form the first condition is not automatic. ω² = 2h₁h₂ is negative on half the plane and zero on the two rulings, which are isotropic, so H_square_pos genuinely selects the positive cone. That is the difference from the rank-one slice, where every nonzero class has positive square, and it makes the quadric the first slice whose Hodge certificate is neither vacuous nor automatic.

                @[simp]

                ∫(f₁+f₂)² = 2, so the Segre class has positive square.

                @[simp]

                The anti-diagonal direction has square -2u²: negative for every nonzero parameter. The ruling classes themselves are isotropic, which is why this needs the anti-diagonal and not an arbitrary transverse direction.

                The quadric's wall slice is Hodge.

                Not vacuous, unlike the rank-one slice: the transverse space is a line and the second clause has to be checked on it.

                The quadric's wall slice is geometric.

                ampleCone and the slice were both already here; nothing had put them together, so IsGeometric had no witness on this surface.

                The Hodge certificate fails off the positive cone, which is what makes the hyperbolic case different from the definite one.

                A ruling class is isotropic, so it cannot be the distinguished direction of any Hodge slice. On a rank-one surface no such class exists.

                @[simp]

                Evaluating the inherited wall family gives the existing arbitrary-real quadric charge in ruling coordinates.

                A nonzero anti-diagonal B-field cannot be represented as beta H for the Segre polarization. This is the concrete information lost by a scalar-only B parameter on a rank-two surface.