Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results The Effect of Dissipation on the Transformation-Based Approximate Electromagnetic Cloaking Scheme Andy Thaler Department of Mathematics University of Utah Math 5440 December 8, 2010 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Outline 1 Helmholtz Equation In 2D, models the propagation of TE and TM waves In 3D, models the propagation of acoustic waves 2 Cloaking for the 2D Helmholtz Equation Boundary Measurements Map Definition of Cloaking Main Idea Behind Transformation-Based Cloaking Dissipation and Results Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Wave Equation Maxwell’s Equations imply that the electric field satisfies the wave equation, namely ∂2E ∇2 E = µ0 0 2 ∂t Assuming the waves are traveling in the y −direction, the solution is of the following form: h i e 0 e i(ky −ωt) E (y , t) = Re E Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Helmholtz Equation ∇ · [A(x)∇u(x)] + ω 2 q(x)u(x) = 0 1 When derived from Maxwell’s Equations, A(x) and q(x) are related to µ(x) and (x). 2 u represents the longitudinal component of the electric field (TM waves) or the magnetic field (TE waves). 3 In 2D, it models the propagation of monochromatic electromagnetic transverse plane waves. 4 In 3D, it models the propagation of acoustic waves. Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results What is Cloaking? 1 2 Assume radial symmetry Cloak against plane waves Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Boundary Measurements Map ∇ · [A(x)∇u(x)] + ω 2 q(x)u(x) = 0 1 2 3 We consider B2 . Maps field to flux for the Helmholtz equation. We used the inverse of this map (more stable numerically). Definition (Boundary Measurements Map) The boundary measurements map associated with the Helmholtz equation is defined by: ΛA,q (u) = (A∇u) · ν , (1) x where ν is the outward normal vector of B2 , and ν(x) = for 2 x ∈ ∂B2 . Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Invariance Principle ∇x · [A(x)∇x u(x)] + ω 2 q(x)u(x) = 0 F : B2 → B2, smooth enough, F(x) = x on ∂B2 . Changing variables with y = F(x) gives ∇y · [F∗ A(y)∇y v (y)] + ω 2 q∗ (y)v (y) = 0 , where v (y) = u[F−1 (y)]. Also, ΛA, q = ΛF∗ A, q∗ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Invariance Principle Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Transformation-Based Cloaking Proposed by Pendry ΛF∗ I ,1∗ = ΛI ,1 A(x), q(x) = F∗ I , 1∗ in B2 \ B1/2 arbitrary real in B1/2 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Transformation-Based Approximate Cloaking Scheme ΛA,q = ΛAρ ,qρ 2ρ y F−1 (y) = ρ 4 (2 − ρ) 2 (2 − ρ) y 2− + |y| 3 3 |y| Andy Thaler 1 for |y| < 2 1 for ≤ |y| ≤ 2 . 2 The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Transformation-Based Approximate Cloaking Scheme ΛA,q = ΛAρ ,qρ 6= ΛI ,1 2ρ y F−1 (y) = ρ 4 (2 − ρ) 2 (2 − ρ) y 2− + |y| 3 3 |y| Andy Thaler 1 for |y| < 2 1 for ≤ |y| ≤ 2 . 2 The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Kohn, Onofrei, Vogelius, and Weinstein 1 2 1 . | log(ρ)| In 3D, error ∼ ρ. In 2D, error ∼ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Dissipation Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Effect of Dissipation Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Effect of Dissipation Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Step 1–Mapping Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Step 2–Setup Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Step 2–Setup qρ (a) ∇2 u1 (x) + ω 2 u1 (x) = 0 Aρ (b) ∇2 u2 (x) + ω 2 (1 + iβ)u2 (x) = 0 (c) ∇2 u3 (x) + ω 2 (1 + iδ)u3 (x) = 0 ∂u1 ∂u2 (d) u1 (ρ, θ) = u2 (ρ, θ); Aρ (ρ, θ) = (ρ, θ) ∂r ∂r ∂u2 ∂u3 (2ρ, θ) = (2ρ, θ) (e) u2 (2ρ, θ) = u3 (2ρ, θ); ∂r ∂r (f) ∂u3 (2, θ) = ψ ∂r Andy Thaler in Bρ in B2ρ \ Bρ in B2 \ B2ρ on ∂Bρ on ∂B2ρ on ∂B2 The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Step 2–Setup where (1) 0 (1) 0 (2ωze ) − J00 (2ωze ) H0 (ωze ρ) ze J0 (ωρ) J00 (ωze ρ) H0 qρ = (1) 0 (1) 0 0 J0 (ωρ) J0 (ωze ρ) H0 (2ωze ) − J0 (2ωze )H0 (ωze ρ) Aρ = qρ I Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Step 2–Setup where (1) 0 (1) 0 (2ωze ) − J00 (2ωze ) H0 (ωze ρ) ze J0 (ωρ) J00 (ωze ρ) H0 qρ = (1) 0 (1) 0 0 J0 (ωρ) J0 (ωze ρ) H0 (2ωze ) − J0 (2ωze )H0 (ωze ρ) Aρ = qρ I ze = 1 + iδ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Step 2–Setup where (1) 0 (1) 0 (2ωze ) − J00 (2ωze ) H0 (ωze ρ) ze J0 (ωρ) J00 (ωze ρ) H0 qρ = (1) 0 (1) 0 0 J0 (ωρ) J0 (ωze ρ) H0 (2ωze ) − J0 (2ωze )H0 (ωze ρ) Aρ = qρ I ze = 1 + iδ δ > 0 denotes the dissipation level in the annulus B2 \ B2ρ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Step 2–Setup where (1) 0 (1) 0 (2ωze ) − J00 (2ωze ) H0 (ωze ρ) ze J0 (ωρ) J00 (ωze ρ) H0 qρ = (1) 0 (1) 0 0 J0 (ωρ) J0 (ωze ρ) H0 (2ωze ) − J0 (2ωze )H0 (ωze ρ) Aρ = qρ I ze = 1 + iδ δ > 0 denotes the dissipation level in the annulus B2 \ B2ρ ψ is the normal derivative of the incoming plane wave on ∂B2 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Separation of Variables for u1 Let u1 (r , θ) = R(r )Θ(θ). Then we have ∇2 u1 + ω 2 qρ u1 = 0 Aρ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Separation of Variables for u1 Let u1 (r , θ) = R(r )Θ(θ). Then we have qρ u1 = 0 Aρ qρ ∂ 2 u1 1 ∂u1 1 ∂ 2 u1 ⇔ + + + ω 2 u1 = 0 2 2 2 ∂r r ∂r r ∂θ Aρ ∇2 u1 + ω 2 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Separation of Variables for u1 Let u1 (r , θ) = R(r )Θ(θ). Then we have qρ u1 = 0 Aρ qρ ∂ 2 u1 1 ∂u1 1 ∂ 2 u1 ⇔ + + + ω 2 u1 = 0 2 2 2 ∂r r ∂r r ∂θ Aρ q 1 1 ρ ⇔ R 00 Θ + R 0 Θ + 2 RΘ00 + ω 2 RΘ = 0 r r Aρ ∇2 u1 + ω 2 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Separation of Variables for u1 Let u1 (r , θ) = R(r )Θ(θ). Then we have qρ u1 = 0 Aρ qρ ∂ 2 u1 1 ∂u1 1 ∂ 2 u1 ⇔ + + + ω 2 u1 = 0 2 2 2 ∂r r ∂r r ∂θ Aρ q 1 1 ρ ⇔ R 00 Θ + R 0 Θ + 2 RΘ00 + ω 2 RΘ = 0 r r Aρ qρ R 00 R 0 Θ00 ⇔ r2 +r + + ω2 r 2 = 0 R R Θ Aρ ∇2 u1 + ω 2 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Separation of Variables for u1 Let u1 (r , θ) = R(r )Θ(θ). Then we have qρ u1 = 0 Aρ qρ ∂ 2 u1 1 ∂u1 1 ∂ 2 u1 ⇔ + + + ω 2 u1 = 0 2 2 2 ∂r r ∂r r ∂θ Aρ q 1 1 ρ ⇔ R 00 Θ + R 0 Θ + 2 RΘ00 + ω 2 RΘ = 0 r r Aρ qρ R 00 R 0 Θ00 ⇔ r2 +r + + ω2 r 2 = 0 R R Θ Aρ qρ R 00 R0 Θ00 ⇔ r2 +r + ω2 r 2 = − =K. R R Aρ Θ ∇2 u1 + ω 2 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic K = 0. Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic K = 0. Θ00 = 0. Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic K = 0. Θ00 = 0. Θ(θ) = C + Dθ. Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic K = 0. Θ00 = 0. Θ(θ) = C + Dθ. 2π−periodic iff D = 0 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic K = 0. Θ00 = 0. Θ(θ) = C + Dθ. 2π−periodic iff D = 0 K > 0 e.g. K = µ2 . Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic K = 0. Θ00 = 0. Θ(θ) = C + Dθ. 2π−periodic iff D = 0 K > 0 e.g. K = µ2 . Θ00 + µ2 Θ = 0. Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic K = 0. Θ00 = 0. Θ(θ) = C + Dθ. 2π−periodic iff D = 0 K > 0 e.g. K = µ2 . Θ00 + µ2 Θ = 0. Θ(θ) = Ee iµθ + Fe −iµθ . Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results 3 Cases K < 0 e.g. K = −µ2 . Θ00 − µ2 Θ = 0. Θ(θ) = Ae µθ + Be −µθ . NOT 2π−periodic K = 0. Θ00 = 0. Θ(θ) = C + Dθ. 2π−periodic iff D = 0 K > 0 e.g. K = µ2 . Θ00 + µ2 Θ = 0. Θ(θ) = Ee iµθ + Fe −iµθ . 2π−periodic iff µ = k, where k is a non-negative integer Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Bessel’s Equation The equation for R becomes 2 2 00 0 2 qρ 2 r − k R = 0. r Rk + rRk + ω Aρ (The boundary conditions will be dealt with momentarily). The solution to Bessel’s Equation of order k is r r qρ qρ (1) Rk (r ) = Pk Jk ω r + Qk H k ω r Aρ Aρ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Superposition Thus u1,k (r , θ) = Rk (r )Θk (θ) r r qρ qρ (1) = Pk Jk ω Ee ikθ + Fe −ikθ . r + Qk H k ω r Aρ Aρ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Superposition Thus u1,k (r , θ) = Rk (r )Θk (θ) r r qρ qρ (1) = Pk Jk ω Ee ikθ + Fe −ikθ . r + Qk H k ω r Aρ Aρ Superposing these solutions gives r r ∞ X qρ qρ (1) r + bk Hk ω r e ikθ + e −ikθ u1 (r , θ) = ak Jk ω Aρ Aρ k=0 r r ∞ X qρ qρ (1) = ak Jk ω r + bk Hk r e ikθ . ω Aρ Aρ k=−∞ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Solution r ∞ X qρ ωr e ikθ ak Jk (a) u1 (r , θ) = A ρ k=−∞ ∞ h i X (1) (b) u (r , θ) = c J (ωz r ) + d H (ωz r ) e ikθ 2 k k b k k b k=−∞ ∞ h i X (1) (c) u (r , θ) = e J (ωz r ) + f H (ωz r ) e ikθ e e 3 k k k k in Bρ in B2ρ \ Bρ in B2 \ B2ρ k=−∞ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Error The error we measured was the following normalized error: R ∂B2 |u3 − u0 R ∂B2 |ψ|2 |2 1 2 1 2 or, equivalently, hR 2π θ=0 |u3 (2, θ) hR − u0 (2, θ) |2 dθ 2π 2 θ=0 |ψ (2, θ) | dθ Andy Thaler i1 2 i1 2 The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Dependence on ρ and δ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Dependence on ρ and δ Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Electromagnetic Wave Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Contour Plot of Electromagnetic Wave Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Shadow Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Cloaking Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Zoom-in Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Bessel’s Equation of Order 0 1 R 00 (r ) + R 0 (r ) + R(r ) = 0 r We assume a series solution of the form R(r ) = ∞ X cn r n n=0 Assuming we can differentiate term-by-term, we have R 0 (r ) = ∞ X ncn r n−1 n=1 R 00 (r ) = ∞ X n(n − 1)cn r n−2 n=2 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Bessel’s Equation of Order 0 Inserting the series into Bessel’s equation gives ∞ X n(n − 1)cn r n=2 ∞ X X 1X cn r n = 0 + ncn r n−1 + r n(n − 1)cn r n−2 + n=2 ∞ X ∞ ∞ n−2 n=1 n=0 ∞ X ∞ X ncn r n−2 + n=1 n(n − 1)cn r n−2 + c1 r −1 + n=2 ∞ X c1 + r Andy Thaler n=0 ncn r n−2 + n=2 cn r n = 0 ∞ X cn−2 r n−2 = 0 n=2 ∞ X n2 cn + cn−2 r n−2 = 0 n=2 The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Bessel’s Equation of Order 0 This implies that c1 = 0 cn = − cn−2 n2 Using induction, we find that the solution to the above recurrence relation is 0 for n odd cn = n (−1)k c0 for n even, where k = 2 2k 2 2 (k!) Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Bessel’s Equation of Order 0 Thus our solution is R(r ) = ∞ X (−1)k c0 k=0 22k = c0 + (k!) ∞ X 2 r 2k (−1)k c0 22k (k!)2 k=0 r 2k The ratio test shows that the above series converges for all r . We thus define ∞ X (−1)k 2k J0 (r ) = r 2k (k!)2 2 k=0 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Bessel’s Equation of Orders 0, 1, 2, and 10 Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Clown Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim Introduction Cloaking for Helmholtz Introduction to Dissipation Solution Results Thank you!!! Also, thank you to Professor Daniel Onofrei. Andy Thaler The Effect of Dissipation on the Transformation-Based Approxim