Poincaré lemma: spatial Hessian data integrate to a potential on
OpenNewtonCartan.exists_potential_of_closed_FLet be a smooth -tensor field on (indices relative to the standard global chart) satisfying
and the closed-Hessian condition
for all spatial indices. Then there exists a smooth scalar potential such that
everywhere. Equivalently, the spatial Hessian of recovers on spatial indices, while the temporal component of the raised gradient vanishes identically.
This is the global Poincaré lemma on the contractible chart for Hessian data; it is the analytic step in Trautman recovery once CurvConds have reduced the Riemann tensor to .
Formalization Note. pd and gradUp are as in Def_NewtonCartan_defs; ContDiff ℝ ∞ F is ContDiff of the uncurried map E → Fin 4 → Fin 4 → ℝ.
DEPRECATED: required full closedness including time, which Trautman F need not satisfy. Replaced by NewtonCartan.exists_potential_of_spatial_F.
import Definitions.Def_NewtonCartan_defs open scoped ContDiff
namespace NewtonCartan
/-- If $F^a{}_d$ is smooth, vanishes whenever $a=0$ or $d=0$, is symmetric in its spatial indices, and is closed as a spatial Hessian ($\partial_k F^i{}_j = \partial_j F^i{}_k$), then there is a smooth $\varphi$ with $\partial_d(\nabla^a\varphi) = F^a{}_d$. -/
theorem exists_potential_of_closed_F (F : E → Fin 4 → Fin 4 → ℝ)
(hF : ContDiff ℝ ∞ F)
(hrow : ∀ x d, F x 0 d = 0) (hcol : ∀ x a, F x a 0 = 0)
(hsym : ∀ x i j, F x i j = F x j i)
(hclosed : ∀ x i j k,
pd (fun y => F y i j) k x = pd (fun y => F y i k) j x) :
∃ φ : E → ℝ, ContDiff ℝ ∞ φ ∧
∀ x a d, pd (fun y => gradUp φ y a) d x = F x a d := by sorry
end NewtonCartan