betaFeedback

WorksPhysics and Astronomy

Closed-form dyadic Green functions for transversely uniaxial bianisotropic medium with equal magnetoelectric scalars

Abstract

A transversely uniaxial bianisotropic medium with equal magnetoelectric scalars was conceived by Weiglhofer in 1994. Using the scalar factorization method, he concluded that the dyadic Green functions of this medium could not be found in closed form. That conclusion is incorrect. The essentials of the derivation are reproduced here, and closed-form expressions are given for all four dyadic Green functions of the unbounded medium, linking the electric and magnetic fields to electric and magnetic source current densities.

Introduction

In 1994, Weiglhofer conceived the transversely uniaxial bianisotropic medium (TUBM) as a special case of the most general uniaxial bianisotropic medium [1]. Such a medium is the homogenized description of a composite of electrically small uniaxial inclusions whose axes lie randomly in a plane normal to a fixed direction [2]. Its constitutive relations D=ε‾‾⋅E+ξ‾‾⋅H\mathbf{D} = \underline{\underline{\varepsilon}} \cdot\mathbf{E} + \underline{\underline{\xi}} \cdot\mathbf{H} and B=ζ‾‾⋅E+μ‾‾⋅H\mathbf{B} = \underline{\underline{\zeta}} \cdot\mathbf{E} + \underline{\underline{\mu}} \cdot\mathbf{H} contain the dyadics

ε‾‾=εtI‾‾+(εz−εt)uzuz,ξ‾‾=ξt(I‾‾−uzuz)ζ‾‾=ζt(I‾‾−uzuz),μ‾‾=μtI‾‾+(μz−μt)uzuz(1)\begin{aligned} \underline{\underline{\varepsilon}} &= \varepsilon_t \underline{\underline{I}} + (\varepsilon_z-\varepsilon_t)\mathbf{u}_z\mathbf{u}_z, \qquad\underline{\underline{\xi}} = \xi_t\left(\underline{\underline{I}}-\mathbf{u}_z\mathbf{u}_z\right) \\ \underline{\underline{\zeta}} &= \zeta_t\left(\underline{\underline{I}}-\mathbf{u}_z\mathbf{u}_z\right), \qquad\underline{\underline{\mu}} = \mu_t \underline{\underline{I}} + (\mu_z-\mu_t)\mathbf{u}_z\mathbf{u}_z \tag*{(1)} \end{aligned}

where ε‾‾\underline{\underline{\varepsilon}} is permittivity dyadic, μ‾‾\underline{\underline{\mu}} is the permeability dyadic, ξ‾‾\underline{\underline{\xi}} and ζ‾‾\underline{\underline{\zeta}} are the magnetoelectric dyadics, and I‾‾\underline{\underline{I}} is the identity dyadic. The scalars ξt\xi_t and ζt\zeta_t capture the magnetoelectric properties of the TUBM. Throughout, the permittivity and permeability dyadics are assumed to be nonsingular, i.e., εtεz≠0\varepsilon_t\varepsilon_z \ne0 and μtμz≠0\mu_t\mu_z \ne0.

Weiglhofer [1] employed the scalar factorization method he developed with Lindell [3, 4, 5] to determine the electric dyadic Green function for the TUBM in closed form. This method requires the factorization of a fourth-order scalar operator H\mathcal{H}. Weiglhofer found that factorization is possible only if

ξt=ζt.(2)\xi_t = \zeta_t . \tag*{(2)}

Despite that factorization, he could not determine the electric dyadic Green function G‾‾ee(r,r′)\underline{\underline{G}}_{ee}(\mathbf{r},\mathbf{r}') in closed form. He conclude that “the transversally uniaxial bianisotropic medium with $\xi_t = \zeta_t becomes the first known case in which a closed-form solution for the dyadic Green function is not forthcoming despite successful factorization of the fourth-order operator H.\mathcal{H}.”

While exploring the capabilities of ChatGPT-6.1 Sol Light to implement the scalar factorization method, we found that Weiglhofer’s conclusion regarding the TUBM with equal magnetoelectric scalars is incorrect. In this short paper the essentials of the derivation are reproduced, and closed-form expressions are given for the dyadic Green functions linking (i) the electric field to the source electric current density, (ii) the electric field to the source magnetic current density, (iii) the magnetic field to the source magnetic current density and (iv) the magnetic field to the source electric current density.

Vectors are in bold and dyadics are double underlined. The unit vectors of the Cartesian coordinate system are ux\mathbf{u}_x, uy\mathbf{u}_y and uz\mathbf{u}_z. The identity dyadic is denoted by I‾‾\underline{\underline{I}} and I‾‾t=I‾‾−uzuz\underline{\underline{I}}_t=\underline{\underline{I}}-\mathbf{u}_z\mathbf{u}_z. The fields are time harmonic, and a factor exp⁡(−iωt)\exp(-i\omega t) is suppressed.

Field equations

With the source electric current density denoted by J\mathbf{J} and the source magnetic current density denoted by K\mathbf{K}, the two Maxwell curl equations are:

(∇×I‾‾−iωζ‾‾)⋅E(r)−iωμ‾‾⋅H(r)=−K(r)iωε‾‾⋅E(r)+(∇×I‾‾+iωξ‾‾)⋅H(r)=J(r)(3)\begin{aligned} (\nabla\times\underline{\underline{I}}-i\omega\underline{\underline{\zeta}})\cdot\mathbf{E}(\mathbf{r})-i\omega\underline{\underline{\mu}}\cdot\mathbf{H}(\mathbf{r})&=-\mathbf{K}(\mathbf{r})\\ i\omega\underline{\underline{\varepsilon}}\cdot\mathbf{E}(\mathbf{r})+(\nabla\times\underline{\underline{I}}+i\omega\underline{\underline{\xi}})\cdot\mathbf{H}(\mathbf{r})&=\mathbf{J}(\mathbf{r}) \tag*{(3)} \end{aligned}

Their solutions in the unbounded medium are written as [6, 7, 8]:

E(r)=iω∭G‾‾ee(r,r′)⋅J(r′) d3r′+∭G‾‾em(r,r′)⋅K(r′) d3r′H(r)=∭G‾‾me(r,r′)⋅J(r′) d3r′+iω∭G‾‾mm(r,r′)⋅K(r′) d3r′(4)\begin{aligned} \mathbf{E}(\mathbf{r})&=i\omega\iiint\underline{\underline{G}}_{ee}(\mathbf{r},\mathbf{r}')\cdot\mathbf{J}(\mathbf{r}')\,\mathrm{d}^3r'+\iiint\underline{\underline{G}}_{em}(\mathbf{r},\mathbf{r}')\cdot\mathbf{K}(\mathbf{r}')\,\mathrm{d}^3r'\\ \mathbf{H}(\mathbf{r})&=\iiint\underline{\underline{G}}_{me}(\mathbf{r},\mathbf{r}')\cdot\mathbf{J}(\mathbf{r}')\,\mathrm{d}^3r'+i\omega\iiint\underline{\underline{G}}_{mm}(\mathbf{r},\mathbf{r}')\cdot\mathbf{K}(\mathbf{r}')\,\mathrm{d}^3r' \tag*{(4)} \end{aligned}

The electric dyadic Green function G‾‾ee\underline{\underline{G}}_{ee} is the same as G‾‾\underline{\underline{G}} of Ref. 1. It satisfies the dyadic differential equation

L‾‾ee⋅G‾‾ee(r,r′)=δ(r−r′)I‾‾,(5)\underline{\underline{L}}_{ee}\cdot\underline{\underline{G}}_{ee}(\mathbf{r},\mathbf{r}')=\delta(\mathbf{r}-\mathbf{r}')\underline{\underline{I}}, \tag*{(5)}

where δ(r−r′)\delta(\mathbf{r}-\mathbf{r}') is the Dirac delta and the dyadic differential operator

L‾‾ee≡(∇×I‾‾+iωξ‾‾)⋅μ‾‾−1⋅(∇×I‾‾−iωζ‾‾)−ω2ε‾‾.(6)\underline{\underline{L}}_{ee}\equiv(\nabla\times\underline{\underline{I}}+i\omega\underline{\underline{\xi}})\cdot\underline{\underline{\mu}}^{-1}\cdot(\nabla\times\underline{\underline{I}}-i\omega\underline{\underline{\zeta}})-\omega^2\underline{\underline{\varepsilon}}. \tag*{(6)}

Essentials of the scalar factorization method

If there exists an adjoint operator L‾‾adj\underline{\underline{L}}_{\mathrm{adj}} defined by

L‾‾adj⋅L‾‾ee=HI‾‾,(7)\underline{\underline{L}}_{\mathrm{adj}}\cdot\underline{\underline{L}}_{ee}=\mathcal{H}\underline{\underline{I}}, \tag*{(7)}

then the solution of Eq. (5) is

G‾‾ee=L‾‾adjW(r,r′),(8)\underline{\underline{G}}_{ee}=\underline{\underline{L}}_{\mathrm{adj}}W(\mathbf{r},\mathbf{r}'), \tag*{(8)}

where the scalar super-Green function W(r,r′)W(\mathbf{r},\mathbf{r}') satisfies the differential equation

HW(r,r′)=δ(r−r′).(9)\mathcal{H}W(\mathbf{r},\mathbf{r}')=\delta(\mathbf{r}-\mathbf{r}'). \tag*{(9)}

For the TUBM with equal magnetoelectric scalars (cf. Eqs. (1) and (2)), the scalar fourth-order operator [1]

H=HεHμ−LεLμ(10)\mathcal{H}=\mathcal{H}_{\varepsilon}\mathcal{H}_{\mu}-\mathcal{L}_{\varepsilon}\mathcal{L}_{\mu} \tag*{(10)}

is composed of

Hε=∇t2+εzμtδt(∂2∂z2+kt2)Hμ=∇t2+εtμzδt(∂2∂z2+kt2)Lε=−εzξtδt(∂2∂z2+kt2)Lμ=−μzξtδt(∂2∂z2+kt2)(11)\begin{aligned} \mathcal{H}_{\varepsilon} &= \nabla_{\mathrm{t}}^2+\frac{\varepsilon_z\mu_t}{\delta_t}\left(\frac{\partial^2}{\partial z^2}+k_t^2\right)\\ \mathcal{H}_{\mu} &= \nabla_{\mathrm{t}}^2+\frac{\varepsilon_t\mu_z}{\delta_t}\left(\frac{\partial^2}{\partial z^2}+k_t^2\right)\\ \mathcal{L}_{\varepsilon} &= -\frac{\varepsilon_z\xi_t}{\delta_t}\left(\frac{\partial^2}{\partial z^2}+k_t^2\right)\\ \mathcal{L}_{\mu} &= -\frac{\mu_z\xi_t}{\delta_t}\left(\frac{\partial^2}{\partial z^2}+k_t^2\right) \tag*{(11)} \end{aligned}

where ∇2≡∇t2+∂2/∂z2\nabla^2 \equiv\nabla_{\mathrm{t}}^2+\partial^2/\partial z^2, δt≡εtμt−ξt2\delta_t \equiv\varepsilon_t\mu_t-\xi_t^2, and kt2=ω2δtk_t^2=\omega^2\delta_t. This operator was factorized by Weiglhofer [1] as

H=H+H−,(12)\mathcal{H}=\mathcal{H}_{+}\mathcal{H}_{-}, \tag*{(12)}

where

H±=∇z2+A±(∂2∂z2+kt2)(13)\mathcal{H}_{\pm}=\nabla_z^2+A_{\pm}\left(\frac{\partial^2}{\partial z^2}+k_t^2\right) \tag*{(13)}

and

A±=12δt{εtμz+εzμt∓[(εtμz−εzμt)2+4εzμzξt2]1/2}.(14)A_{\pm}=\frac{1}{2\delta_t}\left\{\varepsilon_t\mu_z+\varepsilon_z\mu_t\mp\left[(\varepsilon_t\mu_z-\varepsilon_z\mu_t)^2+4\varepsilon_z\mu_z\xi_t^2\right]^{1/2}\right\}. \tag*{(14)}

The square root in A±A_{\pm} is taken on the principal branch. Its real part is nonnegative and, when that real part vanishes, its imaginary part is nonnegative. Reversing its sign merely interchanges the labels ++ and −-. We assume A+≠A−A_{+}\ne A_{-} throughout.

The factorization (12) of H\mathcal{H} is crucial for the initiation of the scalar factorization method [3, 4, 5]. Given Eq. (9), W(r,r′)W(\mathbf{r},\mathbf{r}') satisfies both of the following relations:

∇t2W(r,r′)=−1A+−A−[A+G+(r,r′)−A−G−(r,r′)](∂2∂z2+kt2)W(r,r′)=1A+−A−[G+(r,r′)−G−(r,r′)](15)\begin{aligned} \nabla_{\mathrm{t}}^2W(\mathbf{r},\mathbf{r}')&=-\frac{1}{A_{+}-A_{-}}\left[A_{+}G_{+}(\mathbf{r},\mathbf{r}')-A_{-}G_{-}(\mathbf{r},\mathbf{r}')\right]\\ \left(\frac{\partial^2}{\partial z^2}+k_t^2\right)W(\mathbf{r},\mathbf{r}')&=\frac{1}{A_{+}-A_{-}}\left[G_{+}(\mathbf{r},\mathbf{r}')-G_{-}(\mathbf{r},\mathbf{r}')\right] \tag*{(15)} \end{aligned}

Here, the scalar Green functions

G±(r,r′)=exp⁡(iktD±)4πD±(16)G_{\pm}(\mathbf{r},\mathbf{r}')=\frac{\exp(ik_tD_{\pm})}{4\pi D_{\pm}} \tag*{(16)}

with

D±=+A±ρ2+(z−z′)2ρ=+(r−r′)2−(z−z′)2(17)\begin{aligned} D_{\pm}&=+\sqrt{A_{\pm}\rho^2+(z-z')^2}\\ \rho&=+\sqrt{(\mathbf{r}-\mathbf{r}')^2-(z-z')^2} \tag*{(17)} \end{aligned}

satisfy the differential equations [9]

H±G±(r,r′)=−δ(r−r′).(18)\mathcal{H}_{\pm}G_{\pm}(\mathbf{r},\mathbf{r}')=-\delta(\mathbf{r}-\mathbf{r}'). \tag*{(18)}

Here ρ\rho is the nonnegative real square root. For the outgoing Green functions with the exp⁡(−iωt)\exp(-i\omega t) convention, the branches of kt=ωδtk_t=\omega\sqrt{\delta_t} and D±D_{\pm} are fixed by continuation from ω>0\omega>0, δt>0\delta_t>0, and A±>0A_{\pm}>0, where kt>0k_t>0 and D±>0D_{\pm}>0 away from the source. For complex constitutive parameters, the continuation is taken with the outgoing limiting-absorption prescription, so that the choices of ktk_t and D±D_{\pm} must not be made independently.

The adjoint operator for the TUBM with equal magnetoelectric scalars can be written as

L‾‾adj(∇)=L=1∇t2+L=2(∂2∂z2+kt2)+L=3,(19)\underline{\underline{L}}_{\mathrm{adj}}(\nabla)=L_{=1}\nabla_{\mathrm{t}}^2+L_{=2}\left(\frac{\partial^2}{\partial z^2}+k_t^2\right)+L_{=3}, \tag*{(19)}

where

L‾‾1=−1kt2[μt(∇∇+ω2εtεzμtεz−1)+ω2ξt2(μzI‾‾t−μtuzuz)](20)\underline{\underline{L}}_1=-\frac{1}{k_t^2}\left[\mu_t\left(\nabla\nabla+\omega^2\varepsilon_t\varepsilon_z\mu_t\varepsilon_z^{-1}\right)+\omega^2\xi_t^2\left(\mu_z\underline{\underline{I}}_t-\mu_t\boldsymbol{u}_z\boldsymbol{u}_z\right)\right] \tag*{(20)}
\begin{aligned} \underline{\underline{L}}_2={}&-\frac{1}{k_t^2}\left\{\mu_z\left(\nabla\nabla+\omega^2\varepsilon_t\varepsilon_z\mu_t\varepsilon_z^{-1}\right)-\omega^2\mu_z\xi_t^2\boldsymbol{u}_z\boldsymbol{u}_z \\ &+i\omega\mu_z\xi_t\left[(\nabla\times\boldsymbol{u}_z)\boldsymbol{u}_z+\boldsymbol{u}_z(\nabla\times\boldsymbol{u}_z)\right]\right\} \tag*{(21)} \end{aligned}

and

L‾‾3=(εzμt2+μzξt2−μzδt)δt(∇×uz)(∇×uz)−iμzξtωδt∂∂z[∇t(∇×uz)+(∇×uz)∇t](22)\underline{\underline{L}}_3=\frac{\left(\varepsilon_z\mu_t^2+\mu_z\xi_t^2-\mu_z\delta_t\right)}{\delta_t}(\nabla\times\boldsymbol{u}_z)(\nabla\times\boldsymbol{u}_z)-\frac{i\mu_z\xi_t}{\omega\delta_t}\frac{\partial}{\partial z}\left[\nabla_t(\nabla\times\boldsymbol{u}_z)+(\nabla\times\boldsymbol{u}_z)\nabla_t\right] \tag*{(22)}

with ∇t=∇−(∂/∂z)uz\nabla_t=\nabla-(\partial/\partial z)\boldsymbol{u}_z.

On making use of Eqs. (7), (15), (19), and (22), we obtain

G‾‾ee(r,r′)=G‾‾+(r,r′)+G‾‾−(r,r′)+F‾‾(r,r′)(23)\underline{\underline{G}}_{ee}(\boldsymbol{r},\boldsymbol{r}')=\underline{\underline{G}}_{+}(\boldsymbol{r},\boldsymbol{r}')+\underline{\underline{G}}_{-}(\boldsymbol{r},\boldsymbol{r}')+\underline{\underline{F}}(\boldsymbol{r},\boldsymbol{r}') \tag*{(23)}

where

G‾‾±(r,r′)=∓[A±L‾‾1−L‾‾2]G‾‾±(r,r′)(A+−A−)(24)\underline{\underline{G}}_{\pm}(\boldsymbol{r},\boldsymbol{r}')=\mp\left[A_{\pm}\underline{\underline{L}}_1-\underline{\underline{L}}_2\right]\frac{\underline{\underline{G}}_{\pm}(\boldsymbol{r},\boldsymbol{r}')}{(A_{+}-A_{-})} \tag*{(24)}

and

F‾‾(r,r′)=L‾‾3W(r,r′)(25)\underline{\underline{F}}(\boldsymbol{r},\boldsymbol{r}')=\underline{\underline{L}}_3W(\boldsymbol{r},\boldsymbol{r}') \tag*{(25)}

Following Ref. 1, we split up F‾‾(r,r′)\underline{\underline{F}}(\boldsymbol{r},\boldsymbol{r}') as

F‾‾(r,r′)=F‾‾1(r,r′)+F‾‾2(r,r′)+F‾‾3(r,r′)(26)\underline{\underline{F}}(\boldsymbol{r},\boldsymbol{r}')=\underline{\underline{F}}_1(\boldsymbol{r},\boldsymbol{r}')+\underline{\underline{F}}_2(\boldsymbol{r},\boldsymbol{r}')+\underline{\underline{F}}_3(\boldsymbol{r},\boldsymbol{r}') \tag*{(26)}

where

F‾‾1(r,r′)=(εzμt2+μzξt2−μzδt)δt(∇×uz)(∇×uz)W(r,r′)(27)\underline{\underline{F}}_1(\boldsymbol{r},\boldsymbol{r}')=\frac{\left(\varepsilon_z\mu_t^2+\mu_z\xi_t^2-\mu_z\delta_t\right)}{\delta_t}(\nabla\times\boldsymbol{u}_z)(\nabla\times\boldsymbol{u}_z)W(\boldsymbol{r},\boldsymbol{r}') \tag*{(27)}
F‾‾2(r,r′)=−iμzξtωδt∂∂z[∇t(∇×uz)+(∇×uz)∇t]W(r,r′)(28)\underline{\underline{F}}_2(\boldsymbol{r},\boldsymbol{r}')=-\frac{i\mu_z\xi_t}{\omega\delta_t}\frac{\partial}{\partial z}\left[\nabla_t(\nabla\times\boldsymbol{u}_z)+(\nabla\times\boldsymbol{u}_z)\nabla_t\right]W(\boldsymbol{r},\boldsymbol{r}') \tag*{(28)}

and

F‾‾3(r,r′)=L‾‾3W(r,r′)−[F‾‾1(r,r′)+F‾‾2(r,r′)](29)\underline{\underline{F}}_3(\boldsymbol{r},\boldsymbol{r}')=\underline{\underline{L}}_3W(\boldsymbol{r},\boldsymbol{r}')-\left[\underline{\underline{F}}_1(\boldsymbol{r},\boldsymbol{r}')+\underline{\underline{F}}_2(\boldsymbol{r},\boldsymbol{r}')\right] \tag*{(29)}

Weiglhofer [1] treated F‾‾3(r,r′)\underline{\underline{F}}_3(\boldsymbol{r},\boldsymbol{r}') as the solution of a fourth-order equation of the type

HF‾‾3=Q‾‾δ(r−r′)(30)\mathcal{H}\underline{\underline{F}}_3=\underline{\underline{Q}}\delta(\boldsymbol{r}-\boldsymbol{r}') \tag*{(30)}

where he had Q‾‾≠0\underline{\underline{Q}}\ne0 erroneously. He could not solve Eq. (30) in closed form, which led him to the conclusion that G‾‾ee(r,r′)\underline{\underline{G}}_{ee}(\boldsymbol{r},\boldsymbol{r}') cannot be determined in closed form. In contrast, Eqs. (20a)–(20c) show that L‾‾3\underline{\underline{L}}_3 consists of exactly the two terms that produce F‾‾1(r,r′)\underline{\underline{F}}_1(\boldsymbol{r},\boldsymbol{r}') and F‾‾2(r,r′)\underline{\underline{F}}_2(\boldsymbol{r},\boldsymbol{r}'). In other words, Q‾‾=0‾‾\underline{\underline{Q}}=\underline{\underline{0}} so that

F‾‾3(r,r′)=0(31)\underline{\underline{F}}_3(\boldsymbol{r},\boldsymbol{r}')=0 \tag*{(31)}

Since every term on the right sides of Eqs. (25a) and (25b) differentiates W(r,r′)W(\boldsymbol{r},\boldsymbol{r}') with respect to the transverse coordinates but W(r,r′)W(\boldsymbol{r},\boldsymbol{r}') depends on ρ\rho and z−z′z-z' only, the first of Eqs. (15) can be integrated once with respect to ρ\rho. With the integration constant fixed by regularity on the axis ρ=0\rho=0, we get

∇tW(r,r′)=uρiτ4πkt(A+−A−)ρ(32)\nabla_tW(\boldsymbol{r},\boldsymbol{r}')=\boldsymbol{u}_{\rho}\frac{i\tau}{4\pi k_t(A_{+}-A_{-})\rho} \tag*{(32)}

where

τ=exp⁡(iktD+)−exp⁡(iktD−)(33)\tau=\exp(ik_tD_{+})-\exp(ik_tD_{-}) \tag*{(33)}

and (uρ,uφ,uz)(\mathbf{u}_{\rho},\mathbf{u}_{\varphi},\mathbf{u}_{z}) are the unit vectors of the cylindrical coordinate system. In addition,

(∇×uz)W(r,r′)=[∇tW(r,r′)]×uz.(34)(\nabla\times\mathbf{u}_{z})W(\mathbf{r},\mathbf{r}')=[\nabla_{t}W(\mathbf{r},\mathbf{r}')]\times\mathbf{u}_{z}. \tag*{(34)}

Hence F‾‾1(r,r′)\underline{\underline{F}}_{1}(\mathbf{r},\mathbf{r}') and F‾‾2(r,r′)\underline{\underline{F}}_{2}(\mathbf{r},\mathbf{r}') are available in closed form, and no part of G‾‾ee(r,r′)\underline{\underline{G}}_{ee}(\mathbf{r},\mathbf{r}') remains undetermined. Of course, all results here hold in the sense of distributions. Furthermore, terms in ρ−2\rho^{-2} are understood through their limits on the axis ρ=0\rho=0 [1].

Dyadic Green functions in closed form

Electric field due to electric current

Equations (23), (26), and (31) deliver

G‾‾ee(r,r′)=G‾‾+(r,r′)+G‾‾−(r,r′)+F‾‾1(r,r′)+F‾‾2(r,r′),(35)\underline{\underline{G}}_{ee}(\mathbf{r},\mathbf{r}')=\underline{\underline{G}}_{+}(\mathbf{r},\mathbf{r}')+\underline{\underline{G}}_{-}(\mathbf{r},\mathbf{r}')+\underline{\underline{F}}_{1}(\mathbf{r},\mathbf{r}')+\underline{\underline{F}}_{2}(\mathbf{r},\mathbf{r}'), \tag*{(35)}

where

G‾‾±(r,r′)=±1kt2(A+−A−){(μtA±−μz)(∇∇+ω2εtεzμtε−1)+ω2ξst2[A±μzI‾‾t+(μz−A±μt)uzuz]−iωμzξst[(∇×uz)uz+uz(∇×uz)]}G‾‾±(r,r′),(36)\begin{aligned} \underline{\underline{G}}_{\pm}(\mathbf{r},\mathbf{r}')={}&\pm\frac{1}{k_t^2(A_+-A_-)} \Bigl\{(\mu_tA_{\pm}-\mu_z)\left(\nabla\nabla+\omega^2\varepsilon_t\varepsilon_z\mu_t\varepsilon^{-1}\right)\\ &+\omega^2\xi_{\mathrm{st}}^2\left[A_{\pm}\mu_z\underline{\underline{I}}_t+(\mu_z-A_{\pm}\mu_t)\mathbf{u}_z\mathbf{u}_z\right]\\ &-i\omega\mu_z\xi_{\mathrm{st}}\left[(\nabla\times\mathbf{u}_z)\mathbf{u}_z+\mathbf{u}_z(\nabla\times\mathbf{u}_z)\right]\Bigr\}\underline{\underline{G}}_{\pm}(\mathbf{r},\mathbf{r}'), \tag*{(36)} \end{aligned}
F‾‾1(r,r′)=εzμt2+μzξst2−μzδtδt(A+−A−)M‾‾(r,r′),(37)\underline{\underline{F}}_{1}(\mathbf{r},\mathbf{r}')= \frac{\varepsilon_z\mu_t^2+\mu_z\xi_{\mathrm{st}}^2-\mu_z\delta_t}{\delta_t(A_+-A_-)} \underline{\underline{M}}(\mathbf{r},\mathbf{r}'), \tag*{(37)}

and

F‾‾2(r,r′)=−iμzξstωδt(A+−A−)∂∂z[uz×M‾‾(r,r′)−M‾‾(r,r′)×uz],(38)\underline{\underline{F}}_{2}(\mathbf{r},\mathbf{r}')= -\frac{i\mu_z\xi_{\mathrm{st}}}{\omega\delta_t(A_+-A_-)} \frac{\partial}{\partial z} \left[\mathbf{u}_z\times\underline{\underline{M}}(\mathbf{r},\mathbf{r}')-\underline{\underline{M}}(\mathbf{r},\mathbf{r}')\times\mathbf{u}_z\right], \tag*{(38)}

with

M‾‾(r,r′)=(I‾‾t−2uφuφ)iτ4πktρ−[A+G‾‾+(r,r′)−A−G‾‾−(r,r′)]uφuφ.(39)\underline{\underline{M}}(\mathbf{r},\mathbf{r}')= (\underline{\underline{I}}_t-2\mathbf{u}_{\varphi}\mathbf{u}_{\varphi}) \frac{i\tau}{4\pi k_t\rho} -[A_+\underline{\underline{G}}_+(\mathbf{r},\mathbf{r}')-A_-\underline{\underline{G}}_-(\mathbf{r},\mathbf{r}')] \mathbf{u}_{\varphi}\mathbf{u}_{\varphi}. \tag*{(39)}

Equations (35)–(39) were the ultimate goal of Weiglhofer [1].

Magnetic field due to electric current

With K=0\mathbf{K}=0, the first of Eqs. (3) gives

G‾‾me=μ‾‾−1⋅(∇×I‾‾−iωζ‾‾)⋅G‾‾ee.(40)\underline{\underline{G}}_{me}=\underline{\underline{\mu}}^{-1}\mathbin{\cdot}(\nabla\times\underline{\underline{I}}-i\omega\underline{\underline{\zeta}})\mathbin{\cdot}\underline{\underline{G}}_{ee}. \tag*{(40)}

Accordingly,

G‾‾me(r,r′)=G‾‾me+(r,r′)+G‾‾me−(r,r′)+F‾‾me(r,r′),(41)\underline{\underline{G}}_{me}(\mathbf{r},\mathbf{r}')=\underline{\underline{G}}_{me+}(\mathbf{r},\mathbf{r}')+\underline{\underline{G}}_{me-}(\mathbf{r},\mathbf{r}')+\underline{\underline{F}}_{me}(\mathbf{r},\mathbf{r}'), \tag*{(41)}

where

G‾‾me±(r,r′)=±1kt2(A+−A−)μ‾‾−1⋅{−iωξt(μtA±−μz)∇t∇+ω2[εzμt(μtA±−μz)+μzξt2A±](∇×I‾‾−iωξtI‾‾t)+ω2[μt(εt−εz)(μtA±−μz)−(μt+μz)ξt2A±](∇×uz)uz−iωμzξt[∂∂z∇uz−uzuz∇2+(∇×uz)(∇×uz)]}G‾‾±(r,r′)(42)\begin{aligned} \underline{\underline{G}}_{me\pm}(\mathbf{r},\mathbf{r}')={}&\pm\frac{1}{k_t^2(A_+-A_-)}\underline{\underline{\mu}}^{-1}\mathbin{\cdot}\Bigl\{-i\omega\xi_t(\mu_tA_{\pm}-\mu_z)\nabla_t\nabla\\ &+\omega^2\left[\varepsilon_z\mu_t(\mu_tA_{\pm}-\mu_z)+\mu_z\xi_t^2A_{\pm}\right](\nabla\times\underline{\underline{I}}-i\omega\xi_t\underline{\underline{I}}_t)\\ &+\omega^2\left[\mu_t(\varepsilon_t-\varepsilon_z)(\mu_tA_{\pm}-\mu_z)-(\mu_t+\mu_z)\xi_t^2A_{\pm}\right](\nabla\times\mathbf{u}_z)\mathbf{u}_z\\ &-i\omega\mu_z\xi_t\left[\frac{\partial}{\partial z}\nabla\mathbf{u}_z-\mathbf{u}_z\mathbf{u}_z\nabla^2+(\nabla\times\mathbf{u}_z)(\nabla\times\mathbf{u}_z)\right]\Bigr\}\underline{\underline{G}}_{\pm}(\mathbf{r},\mathbf{r}') \tag*{(42)} \end{aligned}

and

F‾me(r,r′)=μ‾−1⋅{εzμt2+μzξt2−μzδtδt[(∂∂z∇−uz∇2)(∇×uz)−iωξt(∇×uz)(∇×uz)]+iμzξtωδt∂∂z[∂∂z(∇×uz)(∇×uz)−(∂∂z∇−uz∇2)∇t]+iωξt[∇t(∇×uz)+(∇×uz)∇t]}W(r,r′).(43)\underline{F}_{me}(\mathbf{r},\mathbf{r}')=\underline{\mu}^{-1}\boldsymbol{\cdot}\{\frac{\varepsilon_z\mu_t^2+\mu_z\xi_t^2-\mu_z\delta_t}{\delta_t}[(\frac{\partial}{\partial z}\nabla-\mathbf{u}_z\nabla^2)(\nabla\times\mathbf{u}_z)-i\omega\xi_t(\nabla\times\mathbf{u}_z)(\nabla\times\mathbf{u}_z)] \\ +\frac{i\mu_z\xi_t}{\omega\delta_t}\frac{\partial}{\partial z}[\frac{\partial}{\partial z}(\nabla\times\mathbf{u}_z)(\nabla\times\mathbf{u}_z)-(\frac{\partial}{\partial z}\nabla-\mathbf{u}_z\nabla^2)\nabla_t] \\ +i\omega\xi_t[\nabla_t(\nabla\times\mathbf{u}_z)+(\nabla\times\mathbf{u}_z)\nabla_t]\}W(\mathbf{r},\mathbf{r}'). \tag*{(43)}

In the expression for F‾me(r,r′)\underline{F}_{me}(\mathbf{r},\mathbf{r}'), W(r,r′)W(\mathbf{r},\mathbf{r}') appears only through ∇tW(r,r′)\nabla_tW(\mathbf{r},\mathbf{r}') and (∇×uz)W(r,r′)(\nabla\times\mathbf{u}_z)W(\mathbf{r},\mathbf{r}'), which are available in closed form from Eqs. (32) and (34).

Magnetic field due to magnetic current

The duality transformation E→H\mathbf{E}\to\mathbf{H}, H→−E\mathbf{H}\to-\mathbf{E}, J→K\mathbf{J}\to\mathbf{K}, K→−J\mathbf{K}\to-\mathbf{J} [9] preserves Eqs. (3) for the TUBM with equal magnetoscalars provided that εt,z↔μt,z\varepsilon_{t,z}\leftrightarrow\mu_{t,z} and ξt→−ξt\xi_t\to-\xi_t. This transformation leaves δt\delta_t, ktk_t, A±A_{\pm}, G±G_{\pm}, τ\tau and WW unchanged. Hence, G‾mm\underline{G}_{mm} is obtained from G‾ee\underline{G}_{ee} as

G‾mm(r,r′)=G‾mm+(r,r′)+G‾mm−(r,r′)+E‾1(r,r′)+E‾2(r,r′),(44)\underline{G}_{mm}(\mathbf{r},\mathbf{r}')=\underline{G}_{mm+}(\mathbf{r},\mathbf{r}')+\underline{G}_{mm-}(\mathbf{r},\mathbf{r}')+\underline{E}_1(\mathbf{r},\mathbf{r}')+\underline{E}_2(\mathbf{r},\mathbf{r}'), \tag*{(44)}

where

G‾mm±(r,r′)=±1kt2(A+−A−){(εtA±−εz)(∇∇+ω2μtμzεtμ−1)+ω2ξt2[A±εzI‾t+(εz−A±εt)uzuz]+iωεzξt[(∇×uz)uz+uz(∇×uz)]}G±(r,r′),(45)\underline{G}_{mm\pm}(\mathbf{r},\mathbf{r}')=\pm\frac{1}{k_t^2(A_+-A_-)}\left\{(\varepsilon_tA_{\pm}-\varepsilon_z)\left(\nabla\nabla+\omega^2\mu_t\mu_z\varepsilon_t\mu^{-1}\right) +\omega^2\xi_t^2\left[A_{\pm}\varepsilon_z\underline{I}_t+(\varepsilon_z-A_{\pm}\varepsilon_t)\mathbf{u}_z\mathbf{u}_z\right] \right. \\ \left.+i\omega\varepsilon_z\xi_t\left[(\nabla\times\mathbf{u}_z)\mathbf{u}_z+\mathbf{u}_z(\nabla\times\mathbf{u}_z)\right]\right\}G_{\pm}(\mathbf{r},\mathbf{r}'), \tag*{(45)}
E‾1(r,r′)=μzεt2+εzξt2−εzδtδt(A+−A−)M‾(r,r′),(46)\underline{E}_1(\mathbf{r},\mathbf{r}')=\frac{\mu_z\varepsilon_t^2+\varepsilon_z\xi_t^2-\varepsilon_z\delta_t}{\delta_t(A_+-A_-)}\underline{M}(\mathbf{r},\mathbf{r}'), \tag*{(46)}

and

E‾2(r,r′)=iεzξtωδt(A+−A−)∂∂z[uz×M‾(r,r′)−M‾(r,r′)×uz].(47)\underline{E}_2(\mathbf{r},\mathbf{r}')=\frac{i\varepsilon_z\xi_t}{\omega\delta_t(A_+-A_-)}\frac{\partial}{\partial z}\left[\mathbf{u}_z\times\underline{M}(\mathbf{r},\mathbf{r}')-\underline{M}(\mathbf{r},\mathbf{r}')\times\mathbf{u}_z\right]. \tag*{(47)}

Electric field due to magnetic current

With J=0\mathbf{J}=0, the second of Eqs. (3) gives G‾em=−ε‾−1⋅(∇×I‾+iωξ‾)⋅G‾mm\underline{G}_{em}=-\underline{\varepsilon}^{-1}\boldsymbol{\cdot}(\nabla\times\underline{I}+i\omega\underline{\xi})\boldsymbol{\cdot}\underline{G}_{mm}, which is also the dual of G‾me\underline{G}_{me}. Hence,

G‾em(r,r′)=G‾em+(r,r′)+G‾em−(r,r′)+F‾em(r,r′),(48)\underline{G}_{em}(\mathbf{r},\mathbf{r}')=\underline{G}_{em+}(\mathbf{r},\mathbf{r}')+\underline{G}_{em-}(\mathbf{r},\mathbf{r}')+\underline{F}_{em}(\mathbf{r},\mathbf{r}'), \tag*{(48)}

where

G‾em±(r,r′)=∓1kt2(A+−A−)ε‾−1⋅{iωξt(εtA±−εz)∇t∇+ω2[εtμz(εtA±−εz)+εzξt2A±](∇×I‾+iωξtI‾t)+ω2[εt(μt−μz)(εtA±−εz)−(εt+εz)ξt2A±](∇×uz)uz+iωεzξt[∂∂z∇uz−uzuz∇2+(∇×uz)(∇×uz)]}G±(r,r′)(49)\underline{G}_{em\pm}(\mathbf{r},\mathbf{r}')=\mp\frac{1}{k_t^2(A_+-A_-)}\underline{\varepsilon}^{-1}\boldsymbol{\cdot}\{i\omega\xi_t(\varepsilon_tA_{\pm}-\varepsilon_z)\nabla_t\nabla +\omega^2[\varepsilon_t\mu_z(\varepsilon_tA_{\pm}-\varepsilon_z)+\varepsilon_z\xi_t^2A_{\pm}](\nabla\times\underline{I}+i\omega\xi_t\underline{I}_t) \\ +\omega^2[\varepsilon_t(\mu_t-\mu_z)(\varepsilon_tA_{\pm}-\varepsilon_z)-(\varepsilon_t+\varepsilon_z)\xi_t^2A_{\pm}](\nabla\times\mathbf{u}_z)\mathbf{u}_z \\ +i\omega\varepsilon_z\xi_t[\frac{\partial}{\partial z}\nabla\mathbf{u}_z-\mathbf{u}_z\mathbf{u}_z\nabla^2+(\nabla\times\mathbf{u}_z)(\nabla\times\mathbf{u}_z)]\}G_{\pm}(\mathbf{r},\mathbf{r}') \tag*{(49)}

and

F‾‾em(r,r′)=−ε−1⋅{μzεt2+εzξt2−εzδtδt[(∂∂z∇−uz∇2)(∇×uz)+iωξt(∇×uz)(∇×uz)]−iεzξtωδt∂∂z[∂∂z(∇×uz)(∇×uz)−(∂∂z∇−uz∇2)∇t]−iωξt[∇t(∇×uz)+(∇×uz)∇t]}W(r,r′).(50)\underline{\underline{F}}_{em}(\mathbf r,\mathbf r')=-\boldsymbol{\varepsilon}^{-1}\cdot\left\{\frac{\mu_z\varepsilon_t^2+\varepsilon_z\xi_t^2-\varepsilon_z\delta_t}{\delta_t}\left[\left(\frac{\partial}{\partial z}\nabla-\mathbf u_z\nabla^2\right)(\nabla\times\mathbf u_z)+i\omega\xi_t(\nabla\times\mathbf u_z)(\nabla\times\mathbf u_z)\right] -\frac{i\varepsilon_z\xi_t}{\omega\delta_t}\frac{\partial}{\partial z}\left[\frac{\partial}{\partial z}(\nabla\times\mathbf u_z)(\nabla\times\mathbf u_z)-\left(\frac{\partial}{\partial z}\nabla-\mathbf u_z\nabla^2\right)\nabla_t\right] -i\omega\xi_t\left[\nabla_t(\nabla\times\mathbf u_z)+(\nabla\times\mathbf u_z)\nabla_t\right]\right\}W(\mathbf r,\mathbf r'). \tag*{(50)}

Concluding remarks

All four dyadic Green functions of the TUBM with equal magnetoelectric scalars are expressed through G±G_{\pm}, τ\tau, and ρ\rho as well as their derivatives, provided δt≠0\delta_t\ne0, A±≠0A_{\pm}\ne0 and A+≠A−A_+\ne A_-. Written as in Eqs. (32c) and (38c), F‾‾2(r,r′)\underline{\underline{F}}_2(\mathbf r,\mathbf r') and E‾‾2(r,r′)\underline{\underline{E}}_2(\mathbf r,\mathbf r') remain valid when the coefficients of F‾‾1(r,r′)\underline{\underline{F}}_1(\mathbf r,\mathbf r') and E‾‾1(r,r′)\underline{\underline{E}}_1(\mathbf r,\mathbf r') vanish. For ξt=0\xi_t=0 the results reduce to those for a uniaxial dielectric-magnetic medium [10].

With this success of the scalar factorization method, which had unfortunately eluded Weiglhofer [1] because of some derivational errors, we expect this method to produce more novel results soon.

Acknowledgments

Both authors were supported by EPSRC (grant number APP68512).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Declaration of AI use

This paper emerged while experimenting with ChatGPT-6.1 Sol Light to implement the scalar factorization method. All equations were typeset by that chatbot. The authors wrote the textual material and verified the equations.

Data availability

All data analyzed in this paper are included in this paper.

References

  1. [1]Weiglhofer, W. S., 1994, Dyadic Green function for unbounded general uniaxial bianisotropic medium. International Journal of Electronics, 77, 105–115.
  2. [2]Weiglhofer, W. S., Lakhtakia, A., and Monzon, J. C., 1993, Maxwell Garnett model for composites of electrically small uniaxial objects. Microwave and Optical Technology Letters, 6, 681–684.
  3. [3]Lindell, I. V., and Weiglhofer, W. S., 1994, Green dyadic for a uniaxial bianisotropic medium. IEEE Transactions on Antennas and Propagation, 42, 1013–1016.DOI
  4. [4]Weiglhofer, W. S., and Lindell, I. V., 1994a, Fields and potentials in general uniaxial bianisotropic media. I. Axial sources. International Journal of Applied Electromagnetics in Materials, 4, 211–220.
  5. [5]Weiglhofer, W. S., and Lindell, I. V., 1994b, Analytic solution for the dyadic Green function of a nonreciprocal uniaxial bianisotropic medium. Archiv für Elektronik und Übertragungstechnik, 48, 116–119.
  6. [6]Tai, C.-T., 1993, Dyadic Green Functions in Electromagnetic Theory, 2nd edn. (Piscataway, NJ, U.S.A.: IEEE Press).
  7. [7]Olyslager, F., and Lindell, I. V., 2002, Electromagnetics and exotic media: A quest for the holy grail. IEEE Antennas & Propagation Magazine, 44(2), 48–58.DOI
  8. [8]Faryad, M., and Lakhtakia, A., 2018, Infinite-Space Dyadic Green Functions in Electromagnetism (San Ramon, CA, U.S.A.: Morgan & Claypool).DOI
  9. [9]Lindell, I. V., 1992, Methods for Electromagnetic Field Analysis (Oxford, U.K.: Clarendon Press).
  10. [10]Weiglhofer, W. S., 1990, Dyadic Green’s functions for general uniaxial media. IEE Proceedings H, 137, 5–10.

Paper details

vlnc:W_qs5c59rj
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

ChatGPT-6.1 Sol Light

View statement
Looking for related papers…
Contents