The second covariant derivative of a section #
Let ∇ be a smooth connection on a vector bundle V and ∇ᵀ a connection on the tangent
bundle. The second covariant derivative of a smooth section σ of V is the tensor
∇²σ(X, Y) = ∇_X (∇_Y σ) - ∇_{∇ᵀ_X Y} σ.
It is linear over functions in both X and Y, so at each point it is a bilinear map from pairs
of tangent vectors to the fibre. This file constructs it with globally smooth extensions of the
tangent vectors and the tensoriality criterion TauCeti.Manifold.eq_of_contMDiff_tensorial.
The Ricci identity, which expresses the antisymmetric part of ∇²σ through curvature and
torsion, is CovariantDerivative.secondCovariantDerivative_sub_swap, in the curvature module
Curvature/SecondCovariantDerivative.lean.
Main definitions and results #
CovariantDerivative.secondCovariantDerivative: the second covariant derivative∇²σat a point, as a bilinear map. It is zero whenσis not smooth.CovariantDerivative.secondCovariantDerivative_apply: its value on a smooth vector field in the second slot,∇²σ(u, Y) = ∇_u (∇_Y σ) - ∇_{∇ᵀ_u Y} σ.CovariantDerivative.eq_secondCovariantDerivative_iff: this value characterizes∇²σ.CovariantDerivative.covariantDerivative_covariantDerivative_smul_section: the Leibniz rule for the iterated covariant derivative of a function multiple of a section, for aC^kconnection with1 ≤ k.CovariantDerivative.secondCovariantDerivative_smul_apply: the Leibniz rule for the second covariant derivative of a function multiple of a section.CovariantDerivative.secondCovariantDerivative_smul_smul_sub: the double commutator of∇²with multiplication by two functionsf,gis(df ⊗ dg + dg ⊗ df) σ, independently of the connections.
References #
- J. M. Lee, Introduction to Riemannian Manifolds, 2nd ed., Springer GTM 176 (2018), Chapter 4 (the total covariant derivative and the second covariant derivative).
Second covariant derivative of a function multiple. For a C^k connection with 1 ≤ k,
a vector field X, a C^k vector field Y, a C^(k + 1) function f and a C^(k + 1) section
σ,
∇_X ∇_Y (f σ) = f ∇_X ∇_Y σ + (X f) ∇_Y σ + (Y f) ∇_X σ + (X (Y f)) σ.
The second covariant derivative ∇²σ(u, v) = ∇_u (∇_Y σ) - ∇_{∇ᵀ_u Y} σ of a smooth section
σ at x, where Y is any smooth vector field with Y x = v. It is bilinear in (u, v), and
is defined to be zero when σ is not smooth.
Instances For
The second covariant derivative of a smooth section, evaluated on a smooth vector field in
its second slot: ∇²σ(u, Y) = ∇_u (∇_Y σ) - ∇_{∇ᵀ_u Y} σ.
The second covariant derivative of a section which is not smooth is zero.
Evaluation on smooth vector fields in the second slot uniquely characterizes the second covariant derivative of a smooth section.
The second covariant derivative is additive in the section.
The second covariant derivative commutes with multiplication of the section by a constant.
The second covariant derivative of the zero section is zero.
Leibniz rule for the second covariant derivative. For smooth f and σ and a smooth
vector field Y,
∇²(f σ)(u, Y) = f ∇²σ(u, Y) + df(u) ∇_Y σ + df(Y) ∇_u σ + (u (Y f) - (∇ᵀ_u Y) f) σ.
The last coefficient is the second covariant derivative of f for ∇ᵀ.
The double commutator of the second covariant derivative with multiplication. For smooth
functions f, g and a smooth section σ,
∇²(f g σ) - f ∇²(g σ) - g ∇²(f σ) + f g ∇²σ = (df ⊗ dg + dg ⊗ df) σ.