Lecture 10 Interior C 2,α estimate for Newtonian potential (continued) Proposition 1 Consider B1 = BR (x0 ) ⊂ B2R (x0 ) = B2 , f ∈ C α (B2R ), where 0 < � α < 1. Let ω(x) = B2 Γ(x − y)f (y)dy, the Newtonian Potential of f in B2 . Then ω ∈ C 2,α (BR (x0 )) and we have estimate �D2 ω�0;BR + Rα |D2 ω |α;BR ≤ C(�f �0;B2R + Rα |f |α;B2R ), where C = C(n, α) is constant. Proof:(continued) � (V I) = (f (x) − f (x)) Dij Γ(x − y)dy B2 \Bδ (ξ) � ≤ |f |C α (x) |x − x|α | Di Γ(x − y)νj dsy | ∂(B2 \Bδ (ξ)) � � α Di Γ(x − y)νj dsy |) ≤ |f |C α (x) δ (| Di Γ(x − y)νj dsy | + | ∂Bδ (ξ) ∂B2 � � 1 1 δ 1 α ≤ c|f |C α (x) δ ( dsy + dsy ) ( ≤ |y − ξ | ≤ |y − x|) n−1 n−1 2 2 ∂B2 |x − y | ∂Bδ (ξ) |x − y| 1 1 nωn (δ)n−1 ) ≤ c|f |C α (x) (δ α n−1 nωn (2R)n−1 + R (δ/2)n−1 ≤ c|f |C α (x) δ α . � (Dij Γ(x − y) − Dij Γ(x − y))(f (y) − f (x))dy (V ) = B2 \Bδ (ξ) � |DDij Γ(ˆ x − y)||x − x||f |C α (x) |x − y |α dy � 1 ≤ cδ |f |C α (x) |x − y|α dy x − y |n+1 |y−ξ|≥δ |ˆ ≤ B2 \Bδ (ξ) Since |x − y | ≤ |x − ξ | + |ξ − y | ≤ δ 3 + |ξ − y | ≤ |ξ − y | 2 2 and |y − ξ | ≤ |y − x| ˆ + |x ˆ − ξ | ≤ |y − x| ˆ + 1 ˆ =⇒ |y − ξ | ≤ |y − x|, 2 1 δ 1 ≤ |y − x| ˆ + |y − ξ | 2 2 We thus get � dy ≤ cδ|f |C α (x) |y − ξ| n+1−α (V ) ≤ cδ |f |C α (x) |y−ξ|≥δ 3R α−2 � ≤ cδ |f |C α (x) r dr ≤ cδ |f |C α (x) δ � δ 3R 1 rn−1 dr rn+1−α 1 ((3R)α−1 − δ α−1 ) α−1 c ≤ |f | α δ α . 1 − α C (x) Combine all the results, we have shown |Dij ω(x) − Dij ω(x)| ≤ C( thus Rα |f (x)| + |f |C α (x) + |f |C α (x) )|x − x|α , Rα |Dij ω(x) − Dij ω(x)| ≤ C(|f (x)| + Rα |f |C α (x) + Rα |f |C α (x) ), |x − x|α i.e. Rα |D2 ω |α;BR ≤ C(�f �C 0 ;B2R + Rα |f |C α (B2R ) ). Since we have already known that |D2 ω|0 ≤ c(�f �C 0 + Rα |f |α ), we finally get �D2 ω�0;BR + Rα |D2 ω |α;BR ≤ C(�f �0;B2R + Rα |f |α;B2R ). � Exercise: 1) Find a continuous function f s.t. Δu = f does not have a C 2 solution. 2) Find g ∈ C 1 and Δu = g but u is not C 2,1 . Interior C 2,α estimates for Poisson’s equation. Application: u ∈ C 2 (B2R (x0 )), Δu = f, f ∈ C α (B2R (x0 )). Theorem 1 �u�C 2,α (BR ) ≤ C (�u�C 0 (B2R ) + �f �C α (B2R ) ). Rα Proof: Since Δ(u − N f ) = 0, where N f is the Newtonian potential of f , thus from the C 2,α estimate of N f we can get the C 2,α of u. � 2 Theorem 2 Let u ∈ C 2 (Ω), Δu = f, f ∈ C α (Ω), then for any Ω� ⊂⊂ Ω, we have �u�C 2,α (Ω� ) ≤ C(�u�C 0 (Ω) + �f �C α (Ω) ). Proof: Take x ∈ Ω� , choose R s.t.B2R ⊂ Ω, then |u|C 2,α (x) ≤ |u|C 2,α (BR (x)) ≤ ≤ C (�u�C 0 ;B2R + |f |C α (B2R ) ) Rα C (�u�C 0 ;Ω + |f |C α (Ω) ). Rα Taking superior over all x and using previous estimate, we get �u�C 2,α (Ω� ) ≤ C(�u�C 0 (Ω) + �f �C α (Ω) ). � Boundary estimate on Newtonian potential: C 2,α estimate up to the bound­ ary for domain with flat boundary portion. Suppose BR ⊂ B2R , with flat boundary portion B1+ ⊂ B2+ . � Lemma 1 Let f ∈ C α (B2+ ), ω(x) = B + Γ(x − y)f (y)dy be the Newtonian potential of 2 f in B2+ . Then ω ∈ C 2,alpha (B1+ ) and |D2 ω |0;B1 + Rα |D2 ω |α;B + ≤ C(|f |0;B2 + Rα |f |α;B + ). 2 1 Proof: Examine the proof form last time. From C 2 estimate have � � Dij ω(x) = Dij Γ(x − y)(f (y) − f (x))dy − f (x) Di Γ(x − y)νj (y)dsy . B2+ (Note � ∂B2+ ∂B2+ Di Γ(x − y)νj (y)dsy = � ∂B2+ Dj Γ(x − y)νi (y)dsy .) So if either i �= n or j = � n, the integral on the lower boundary portion of B2+ vanishes. (Since ν = (0, 0, · · · , −1).) If i = j = n, then � Dnn ω = f (x) (Dn Γ(x − y) − Dn Γ(x − y))(−1)dσ ∂B2+ � ≤ |f (x)| + ∂B2 |DDn Γ(ˆ x − y)||x − x|dσ � ≤ |f (x)|δ ∂B2+ 1 dσ |ˆ x − y |n 1 nωRn−1 Rn ≤ c|f (x)|δ α . ≤ |f (x)|δ Since we know Δω = f , thus ωnn = f − ω11 − ω22 − · · · , so we see that ωnn ∈ C α , and we can get estimate for ωnn from estimates for ωii , i < n. � 3