Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Corollary 2.27 — ∂ᵢⱼu(x) = ∫_{B_R(x)} ∂ᵢⱼΓ(x−y)[f(y)−f(x)] dy − (1/n) f(x) δᵢⱼ

Proved
HunterPDE.Newtonian.secondPartial_newtonianPotential_ball

by mikedeng1 · Sep 27, 2026 · Mathlib 0df444a (Lean v4.33.1)

newtonian-potentialp2o-batch-books5p2o-gran-per-chapterp2o-plan-bookp2o-v1pdesingular-integrals

Let n≥2n \ge 2n≥2, f∈Cc∞(Rn)f \in C_c^\infty(\mathbb{R}^n)f∈Cc∞​(Rn) and u=Γ∗fu = \Gamma * fu=Γ∗f. Let x∈Rnx \in \mathbb{R}^nx∈Rn and let BR(x)B_R(x)BR​(x) be an open ball centered at xxx containing the support of fff. Then for all indices i,ji, ji,j

∂iju(x)=∫BR(x)∂ijΓ(x−y) [f(y)−f(x)] dy−1nf(x) δij,\partial_{ij} u(x) = \int_{B_R(x)} \partial_{ij}\Gamma(x - y)\,\bigl[f(y) - f(x)\bigr]\,dy - \frac{1}{n} f(x)\,\delta_{ij},∂ij​u(x)=∫BR​(x)​∂ij​Γ(x−y)[f(y)−f(x)]dy−n1​f(x)δij​,

where ∂ijΓ(x−y)\partial_{ij}\Gamma(x - y)∂ij​Γ(x−y) is the second partial derivative of Γ\GammaΓ evaluated at x−yx - yx−y and δij\delta_{ij}δij​ is the Kronecker delta. The kernel ∂ijΓ\partial_{ij}\Gamma∂ij​Γ is not locally integrable, and the subtraction of f(x)f(x)f(x) is what makes the integral converge; this representation is the input to the Hölder estimate, Theorem 2.28.

Formalization Note. Indices are 0-based. The statement asserts, as a first conjunct, that y↦∂ijΓ(x−y)[f(y)−f(x)]y \mapsto \partial_{ij}\Gamma(x-y)[f(y)-f(x)]y↦∂ij​Γ(x−y)[f(y)−f(x)] is Lebesgue integrable on BR(x)B_R(x)BR​(x) (an absolutely convergent integral, as the book's proof uses), and then the identity. The support is the closed support tsupport f, and R>0R > 0R>0.

Preamble
import Mathlib
import Definitions.Def_HunterPDE_Newtonian_FundamentalSolution
import Definitions.Def_HunterPDE_Newtonian_NewtonianPotential
import Definitions.Def_HunterPDE_Newtonian_PartialDeriv
Formal statement
namespace HunterPDE.Newtonian

open MeasureTheory
open scoped ContDiff

/-- Hunter, *Notes on PDEs*, p. 39, Corollary 2.27: if `f ∈ C_c^∞(ℝⁿ)` (`n ≥ 2`) and
`u = Γ ∗ f`, then for every `x` and every open ball `B_R(x)` centered at `x` that contains the
support of `f`,
`∂ᵢⱼu(x) = ∫_{B_R(x)} ∂ᵢⱼΓ(x − y) [f(y) − f(x)] dy − (1/n) f(x) δᵢⱼ`.
Here `∂ᵢⱼΓ(x − y)` is the second partial derivative of `Γ` evaluated at `x − y`; the integrand
is Lebesgue integrable on the ball (asserted as the first conjunct, as the proof of Theorem 2.26
states). Indices are 0-based. -/
theorem secondPartial_newtonianPotential_ball (n : ℕ) (hn : 2 ≤ n)
    (f : EuclideanSpace ℝ (Fin n) → ℝ) (hf : ContDiff ℝ ∞ f) (hfc : HasCompactSupport f)
    (x : EuclideanSpace ℝ (Fin n)) (R : ℝ) (hR : 0 < R)
    (hsupp : tsupport f ⊆ Metric.ball x R) (i j : Fin n) :
    IntegrableOn (fun y => secondPartial (fundamentalSolution n) i j (x - y) * (f y - f x))
        (Metric.ball x R) ∧
      secondPartial (newtonianPotential n f) i j x =
        (∫ y in Metric.ball x R, secondPartial (fundamentalSolution n) i j (x - y) * (f y - f x))
          - 1 / (n : ℝ) * f x * (if i = j then 1 else 0) := by sorry

end HunterPDE.Newtonian
Source
Hunter, Notes on Partial Differential Equations (revised 6/18/2014), p. 39, Corollary 2.27, Eq. (2.29)
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Setting.

  • nnn is a natural number with n≥2n \ge 2n≥2.
  • f:Rn→Rf : \mathbb{R}^n \to \mathbb{R}f:Rn→R is infinitely differentiable (C∞C^\inftyC∞) with compact support.
  • x∈Rnx \in \mathbb{R}^nx∈Rn is a point.
  • RRR is a real number with R>0R > 0R>0.
  • The closed support {f≠0}‾\overline{\{f \ne 0\}}{f=0}​ is contained in the open ball BR(x)={y:∣y−x∣<R}B_R(x) = \{y : |y - x| < R\}BR​(x)={y:∣y−x∣<R}.
  • i,j∈{0,…,n−1}i, j \in \{0, \dots, n-1\}i,j∈{0,…,n−1} are coordinate indices.

Symbols.

  • Γn\Gamma_nΓn​ is:

    • for n=2n = 2n=2, Γ2(z)=−12πlog⁡∣z∣\Gamma_2(z) = -\frac{1}{2\pi}\log|z|Γ2​(z)=−2π1​log∣z∣;
    • for n≥3n \ge 3n≥3, Γn(z)=1n(n−2)αn∣z∣−(n−2)\Gamma_n(z) = \frac{1}{n(n-2)\alpha_n}|z|^{-(n-2)}Γn​(z)=n(n−2)αn​1​∣z∣−(n−2);

    where αn\alpha_nαn​ is the Lebesgue measure of the open unit ball, and Γn(0)=0\Gamma_n(0) = 0Γn​(0)=0.

  • u(z)=∫RnΓn(z−y)f(y) dyu(z) = \int_{\mathbb{R}^n}\Gamma_n(z-y)f(y)\,dyu(z)=∫Rn​Γn​(z−y)f(y)dy is the Newtonian potential. It is a Lebesgue integral, taken to be 000 where the integrand is not integrable.

  • For any function ggg:

    • ∂jg(z)\partial_j g(z)∂j​g(z) is the Fréchet derivative of ggg at zzz applied to the standard basis vector eje_jej​, and is 000 where ggg is not differentiable.
    • ∂ijg(z)=∂i(∂jg)(z)\partial_{ij} g(z) = \partial_i(\partial_j g)(z)∂ij​g(z)=∂i​(∂j​g)(z): differentiate in direction jjj first, then in direction iii. The outer derivative is taken of the everywhere-defined function ∂jg\partial_j g∂j​g, including its zero junk values.

Assertion. The statement asserts both of the following.

  1. Integrability. The function y↦∂ijΓn(x−y) (f(y)−f(x))y \mapsto \partial_{ij}\Gamma_n(x-y)\,\big(f(y) - f(x)\big)y↦∂ij​Γn​(x−y)(f(y)−f(x)) is Lebesgue integrable on the ball BR(x)B_R(x)BR​(x).

  2. Formula.

∂iju(x)=∫BR(x)∂ijΓn(x−y) (f(y)−f(x)) dy  −  1n f(x) δij,\partial_{ij} u(x) = \int_{B_R(x)} \partial_{ij}\Gamma_n(x-y)\,\big(f(y)-f(x)\big)\,dy \;-\; \frac{1}{n}\, f(x)\,\delta_{ij},∂ij​u(x)=∫BR​(x)​∂ij​Γn​(x−y)(f(y)−f(x))dy−n1​f(x)δij​,

where δij\delta_{ij}δij​ is 111 if i=ji = ji=j and 000 otherwise.

Degenerate cases:

  • Excluded dimensions. n≥2n \ge 2n≥2 excludes n=0,1n = 0, 1n=0,1.
  • The point y=xy = xy=x. In the integrand, ∂ijΓn(x−y)\partial_{ij}\Gamma_n(x - y)∂ij​Γn​(x−y) at y=xy = xy=x is the value produced by the derivative conventions at 000, possibly the junk value 000. This is a single point of measure zero.
  • Satisfiability of the support hypothesis. For any such fff and any xxx, a sufficiently large RRR satisfies it. If f≡0f \equiv 0f≡0, every R>0R > 0R>0 works, and both sides of the formula are 000.
  • Choice of RRR. The statement quantifies over every admissible RRR. So it asserts that the right-hand side takes the same value for all such radii.
  • Symmetry. The statement is made separately for each ordered pair (i,j)(i, j)(i,j). No symmetry of ∂ij\partial_{ij}∂ij​ is assumed.
Human review
  • Endorsed by Shuze Chen · Oct 2, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 2, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me