Validity limits and spatial-dispersion-aware generalization of local effective-parameter retrieval for resonant metamaterials

Koschny, Markoš, Smith, and Soukoulis (Phys. Rev. E 68, 065602 (2003)) found that inverting the reflection and transmission of a metamaterial slab under the homogeneous-medium assumption yields antiresonant behavior — the poles of one retrieved parameter coincide with the zeros of the other — and imaginary parts of the effective permittivity and permeability of opposite sign near the magnetic resonance of split-ring-resonator arrays, and explained both features by the boundedness of the refractive index imposed by finite lattice periodicity. The same paper acknowledged that this local, frequency-only description is only approximate, because spatial dispersion is always present and the internal wavelength becomes comparable to the lattice period near resonance, so homogeneous-slab retrieval may lose meaning there. This task asks for a physically meaningful characterization of finite-periodicity resonant metamaterials: quantitative criteria delimiting when locally retrieved $\epsilon(\omega)$ and $\mu(\omega)$ are trustworthy, retrieval formulations that incorporate spatially dispersive (nonlocal) constitutive relations where they are not, and a passivity- and energy-consistent interpretation of the antiresonant features. The sign/passivity aspect has since been largely resolved; the quantitative validity criterion and the general spatial-dispersion-aware retrieval framework remain open.

Background

Negative-index (double-negative) metamaterials are artificial composites whose unit cells — typically split-ring resonators (SRRs, small metallic rings with a gap whose circulating currents produce a resonant magnetic response) combined with metallic wire arrays that act as dilute plasmas — are engineered so that an effective permittivity $\epsilon(\omega)$ and permeability $\mu(\omega)$ are simultaneously negative over a band. The standard characterization tool is parameter retrieval (Smith, Schultz, Markoš, and Soukoulis, Phys. Rev. B 65, 195104 (2002)): the complex reflection and transmission coefficients computed or measured for a slab of the composite are inverted using the textbook formulas for a homogeneous slab to obtain the refractive index $n$ and impedance $z$, from which $\epsilon=n/z$ and $\mu=nz$ follow; the procedure assumes the composite behaves as a continuous medium.

Koschny, Markoš, Smith, and Soukoulis (2003) analyzed what this inversion produces near resonance. For SRR arrays the resonant $\mu(\omega)$ is accompanied by an antiresonant $\epsilon(\omega)$, and $\epsilon''$ and $\mu''$ (imaginary parts, with time dependence $\exp(-i\omega t)$) take opposite signs in part of the resonance band, seemingly contradicting the naive expectation that both are positive in a passive medium. They explained both features through the bounded refractive index: a lattice with period $a$ supports a maximum wave number $k_{\max}=\pi/a$, so the index cannot exceed roughly $n_{\max}\sim c\pi/(\omega_0 a)$; if $n(\omega)=\sqrt{x(\omega)y(\omega)}$ must stay bounded while the resonant parameter $x$ passes through a pole, the second parameter $y$ is forced through the corresponding zero, with opposite-sign imaginary parts following. In arrays of periodically cut wires the roles of $\epsilon$ and $\mu$ are simply exchanged, confirming that the effect is intrinsic to finite periodicity rather than to any particular resonator.

The authors' stated caveat is the origin of the problem addressed here: spatial dispersion — dependence of the constitutive response on the wave vector as well as the frequency, so that the polarization at a point is not determined by the local field alone — is always present to some degree in composites, so a description by bulk, local $\epsilon(\omega)$ and $\mu(\omega)$ is only approximate, and homogeneous-slab retrieval might fail near resonance, where the wavelength inside the composite becomes comparable with the lattice period. The debate that followed also sharpened the observable side of the issue: whether negative $\epsilon''$ or $\mu''$ by itself violates passivity, and how electromagnetic energy should be defined and apportioned in a lossy, dispersive, spatially dispersive medium (Efros 2004; Tretyakov 2005). Modern homogenization research works with Bloch modes (the eigenmodes of the periodic structure) and nonlocal constitutive relations, in which the polarization depends on the field over the unit cell or on wave vector, and these are the natural vocabulary for any answer.

Problem Statement

Establish a physically meaningful effective-medium characterization of finite-periodicity resonant metamaterials, replacing the purely local, homogeneous-slab picture whose failure signs Koschny et al. (2003) identified. Concretely, with $a$ the lattice period, $\lambda_\mathrm{in}$ the wavelength of the Bloch mode inside the composite, and $\epsilon(\omega)$, $\mu(\omega)$ the retrieved local parameters, determine: (i) quantitative criteria — in terms of $a/\lambda_\mathrm{in}$, resonance strength, loss, incidence direction, and the content of higher-order Bloch modes — delimiting the frequency range in which inversion of slab reflection and transmission under the homogeneous-medium assumption yields physically meaningful local $\epsilon(\omega)$ and $\mu(\omega)$ (local parameters that, over that range, reproduce the slab's reflection and transmission consistently as the slab thickness and incidence angle are varied and that satisfy the passivity and causality constraints an effective medium must obey), generalizing the paper's qualitative observation that retrieval fails where the internal wavelength becomes comparable to the period into a criterion that holds for finite-periodicity resonant metamaterials generally, across unit-cell geometries and resonator types and not only for the specific structures the paper examined; (ii) retrieval and homogenization formulas that account for nonlocal (spatially dispersive) constitutive relations in the regime where local retrieval fails, yielding effective parameters that remain predictive off-design (for example, at oblique incidence or for different slab thicknesses) and physically interpretable near the resonance, rather than exhibiting unexplained antiresonant pole–zero inversions; and (iii) the passivity and energy interpretation of the antiresonant features — the opposite signs of $\epsilon''$ and $\mu''$ — connecting them to observable quantities such as stored and dissipated energy densities. Items (i)–(iii) serve one common objective, a validated characterization framework for resonant metamaterials. The sign/passivity part of item (iii) — negative $\epsilon''$ or $\mu''$ does not by itself violate passivity in spatially dispersive media — is settled in the literature and need not be re-derived; the part of item (iii) that remains open is the energy interpretation, connecting the opposite signs to observable stored and dissipated energy densities.

References

1. T. Koschny, P. Markoš, D. R. Smith, C. M. Soukoulis, Resonant and antiresonant frequency dependence of the effective parameters of metamaterials, Phys. Rev. E 68, 065602 (2003). DOI: 10.1103/PhysRevE.68.065602, https://doi.org/10.1103/PhysRevE.68.065602
2. D. R. Smith, S. Schultz, P. Markoš, C. M. Soukoulis, Determination of effective permittivity and permeability of metamaterials from reflection and transmission coefficients, Phys. Rev. B 65, 195104 (2002). DOI: 10.1103/PhysRevB.65.195104, https://doi.org/10.1103/PhysRevB.65.195104
3. Ricardo A. Depine and Akhlesh Lakhtakia, Comment I on "Resonant and antiresonant frequency dependence of the effective parameters of metamaterials", Phys. Rev. E 70, 048601 (2004). DOI: 10.1103/PhysRevE.70.048601, https://doi.org/10.1103/PhysRevE.70.048601
4. A. L. Efros, Comment II on "Resonant and antiresonant frequency dependence of the effective parameters of metamaterials", Phys. Rev. E 70, 048602 (2004). DOI: 10.1103/PhysRevE.70.048602, https://doi.org/10.1103/PhysRevE.70.048602
5. T. Koschny, P. Markoš, D. R. Smith, C. M. Soukoulis, Reply to Comments on "Resonant and antiresonant frequency dependence of the effective parameters of metamaterials", Phys. Rev. E 70, 048603 (2004). DOI: 10.1103/PhysRevE.70.048603, https://doi.org/10.1103/PhysRevE.70.048603
6. Divitha Seetharamdoo, Ronan Sauleau, Kouroch Mahdjoubi, Anne-Claude Tarot, Effective parameters of resonant negative refractive index metamaterials: Interpretation and validity, J. Appl. Phys. 98, 063505 (2005). DOI: 10.1063/1.2041841, https://doi.org/10.1063/1.2041841
7. S. A. Tretyakov, Electromagnetic Field Energy Density in Dispersive and Lossy Metamaterials, in Electromagnetic Materials: Proceedings of the Symposium R, World Scientific, pp. 10–17 (2005). DOI: 10.1142/9789812701718_0002, https://doi.org/10.1142/9789812701718_0002
8. David R. Smith and John B. Pendry, Homogenization of metamaterials by field averaging (invited paper), J. Opt. Soc. Am. B 23, 391 (2006). DOI: 10.1364/JOSAB.23.000391, https://doi.org/10.1364/JOSAB.23.000391
9. Vadim A. Markel, Can the imaginary part of the permeability be negative?, Phys. Rev. E 78, 026608 (2008). DOI: 10.1103/PhysRevE.78.026608, https://doi.org/10.1103/PhysRevE.78.026608
10. Andrea Alù, First-principles homogenization theory for periodic metamaterials, Phys. Rev. B 84, 075153 (2011). DOI: 10.1103/PhysRevB.84.075153, https://doi.org/10.1103/PhysRevB.84.075153
11. Andrea Alù, Arthur D. Yaghjian, Robert A. Shore, Mário G. Silveirinha, Causality relations in the homogenization of metamaterials, Phys. Rev. B 84, 054305 (2011). DOI: 10.1103/PhysRevB.84.054305, https://doi.org/10.1103/PhysRevB.84.054305
12. C. R. Simovski, On electromagnetic characterization and homogenization of nanostructured metamaterials, J. Opt. 13, 013001 (2011). DOI: 10.1088/2040-8978/13/1/013001, https://doi.org/10.1088/2040-8978/13/1/013001
13. Andrei Andryieuski, Sangwoo Ha, Andrey A. Sukhorukov, Yuri S. Kivshar, Andrei V. Lavrinenko, Bloch-mode analysis for retrieving effective parameters of metamaterials, Phys. Rev. B 86, 035127 (2012). DOI: 10.1103/PhysRevB.86.035127, https://doi.org/10.1103/PhysRevB.86.035127
14. Igor Tsukerman and Vadim Markel, A non-asymptotic homogenization theory for periodic electromagnetic structures, Proc. R. Soc. A 470, 20140245 (2014). DOI: 10.1098/rspa.2014.0245, https://doi.org/10.1098/rspa.2014.0245
15. Babak Alavikia, Ali Kabiri, Omar M. Ramahi, Poynting theorem constraints on the signs of the imaginary parts of the electromagnetic constitutive parameters, J. Opt. Soc. Am. A 32, 522 (2015). DOI: 10.1364/JOSAA.32.000522, https://doi.org/10.1364/JOSAA.32.000522
16. Achiles F. Mota, Augusto Martins, John Weiner, Fernando L. Teixeira, Ben-Hur V. Borges, Constitutive parameter retrieval for uniaxial metamaterials with spatial dispersion, Phys. Rev. B 94, 115410 (2016). DOI: 10.1103/PhysRevB.94.115410, https://doi.org/10.1103/PhysRevB.94.115410
17. Karim Mnasri, Andrii Khrabustovskyi, Michael Plum, Carsten Rockstuhl, Retrieving effective material parameters of metamaterials characterized by nonlocal constitutive relations, Phys. Rev. B 99, 035442 (2019). DOI: 10.1103/PhysRevB.99.035442, https://doi.org/10.1103/PhysRevB.99.035442
18. SeokJae Yoo, Suyeon Lee, Jong-Ho Choe, Q-Han Park, Causal homogenization of metamaterials, Nanophotonics 8, 1063 (2019). DOI: 10.1515/nanoph-2019-0069, https://doi.org/10.1515/nanoph-2019-0069
19. Boris Gralak, Spatial dispersion and the sign of the imaginary part of the permeability, recorded research talk on Cassyni (2022). DOI: 10.52843/cassyni.yvpc61, https://doi.org/10.52843/cassyni.yvpc61
20. Ana C. Escobar, Francisco Mesa, Oscar Quevedo-Teruel, Juan D. Baena, Homogenization of Periodic Structures Using the Multimodal Transfer Matrix Method, IEEE Trans. Antennas Propag. 71, 4976 (2023). DOI: 10.1109/TAP.2023.3267176, https://doi.org/10.1109/TAP.2023.3267176