https://doi.org/10.1051/epjam/2016007
Research Article
Extraction of material parameters in 1D fractal metamaterial
Instituto de Telecomunicações, Departamento de Eletrónica, Telecomunicações e Informática, Universidade de Aveiro, 3810193
Aveiro, Portugal
^{*} email: samaneh@ua.pt
^{**} email: jcpedro@ua.pt
Received:
15
April
2016
Accepted:
19
June
2016
Published online: 9 August 2016
In this paper, we develop a transfer matrixbased homogenization approach applicable to selfsimilar fractal structures. With this method, we derive the material parameters of a layered metamaterial comprising two different dielectrics with permittivities equal to ε_{1} and ε_{2}, respectively, in which the layers are distributed based on a fractal set. The band dispersion diagram and effective permittivity obtained analytically are verified with numerical simulations.
Key words: Metamaterials / Fractals / Material parameters
© S. Moeini & J.C. Pedro, Published by EDP Sciences, 2016
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
1 Introduction to the problem
The complexity of metamaterials (MTMs) [1] gives rise to various homogenization techniques, which allow for an efficient analytical description of their electromagnetic properties without the need to consider all of their structural details [2–7].
The majority of the known homogenization methods for MTMs rely on the periodicity of the structure. Although the spatial period in such MTMs is typically much smaller than the wavelength, it is always finite, and thus a finitesize MTM sample always contains a finite number of unit cells.
In this work, we are interested in fractal [8] MTMs, constituents of which follow a fractal downscaling rule and therefore have (theoretically) infinitely many internal parts of infinitesimal dimensions. Here, the selfsimilarity properties of fractals are exploited for proposing a new homogenization method for such MTMs based on fractal geometry.
As a test problem for this homogenization method, we consider a layered fractal MTM comprising two different dielectrics. In this case, an unbounded fractal MTM can be formed by periodically repeating fractal unit cells. In the following we consider a structure having just one such cell. Figure 1 depicts examples of the unit cells with increasing fractal order. In the first order (Figure 1a), we consider three layers: one layer of ε_{1} with the thickness of d_{1} in the middle and two layers with permittivity ε_{2} and the thickness of d_{2}. The structure is along the z axis and has a total thickness of L = d_{1} + 2d_{2}. The layers are considered infinite along y and x axes (as compared to the wavelength).
Figure 1. ε profile for different orders of fractal MTM. In this example, L = d_{1} + 2 × d_{2} = 1, d_{2} =0.45L, ε_{1} = 4 and ε_{2} = 1. (a) 1st order, (b) 2nd order, (c) 3rd order, (d) 4th order. 
By increasing the fractal order, the first and third layers of the firstorder structure are replaced by a triplet of layers which are selfsimilar to the original structure, but with the roles of ε_{1} and ε_{2} interchanged. For a fractal of nth order this procedure is repeated n − 1 times. As an example, the first four resulting profiles ε(z) are shown in Figure 1.
The geometry of the fractal structure is determined by one parameter: the fractal ratio, r, which is defined as and thus by definition can not be higher than 0.5.
2 Analytical transfer matrixbased approach
The effective parameters of the 1D fractal medium can be calculated analytically using the transfer (ABCD) matrix approach. The transfer matrix, A, is defined as follows [9]:(1)in which and are the transverse fields at the input and the output of the structure and A_{diel} is the transfer matrix of a dielectric layer which can be written as shown below [10]:(2)where d is the thickness of the dielectric layer, and(3) (4)and η_{0} and k_{0} are the wave impedance and the wave propagation factor in free space, respectively.
For nonmagnetic materials, we have μ_{r} = 1 and if we use the field units such that η_{0} = 1, the equation for A_{diel} with relative permittivity ε_{r} can be simplified as shown below:(5)
Using these expressions, the total transfer matrix of a stack of dielectric layers is calculated as an ordered product of the transfer matrices of the separate layers.
2.1 Eigenvalues of the transfer matrix and the band dispersion diagram
By knowing the eigenvalues of the transfer matrix, the band diagram [11, p. 29] of the structure can be calculated. The eigenvalues Λ_{1,2} of the total transfer matrix are related to the propagation factors as . Within propagation bands, k_{z} is real (when there is no loss) and the two eigenvalues of the transfer matrix are complex conjugate of each other. On the other hand, within stopbands, k_{z} is purely imaginary and the two eigenvalues have distinct real values.
The band diagram (the dispersion characteristic) for the fractal structure of 9th order obtained with the transfer matrix approach is shown in Figure 2. In this diagram, the propagation factor k_{z} is calculated as:(6)
Figure 2. The band diagram for the 9th order fractal MTM. Four propagation bands and three band gaps are seen in the diagram. 
2.2 Effective permittivity in the quasistatic approximation
When the fractal order is increasing, the total number of layers in the structure grows very fast (in geometric progression), which makes direct numerical computation of the total transfer matrix and the band diagram inefficient. Therefore, to describe fractal structures of order n → ∞, an alternative approach is needed.
In order to develop such approach, we note that the transfer matrix of the whole structure of infinite fractal order A_{∞} can be calculated recursively due to the selfrepeating property of the fractal:(7)where B_{∞} is the transfer matrix of the two fractal parts which surround the middle dielectric layer of thickness d_{1} = (1−2r)L. By using fractal’s selfsimilarity property, B_{∞} can be expressed as(8)
The equations (7) and (8) are functional equations for the unknowns A_{∞} and B_{∞}. Resolving these equations analytically for arbitrary values of the parameters appears impossible. However, when interested only in the behaviour of the structure at low frequencies, the following method can be used. The same approach also allows us to find the effective permittivities of the structure at low frequencies.
In order to solve equations (7) and (8) with the effective medium approach, we equate the unknown transfer matrices A_{∞} and B_{∞} to the transfer matrices of uniform dielectric layers with some effective permittivities ε_{A,eff} and ε_{B,eff}: A_{∞} = A_{diel}(k_{0}, L, ε_{A,eff}) and B_{∞} = A_{diel}(k_{0}, rL, ε_{B,eff}), and search for such ε_{A,eff} and ε_{B,eff} which satisfy the equations (7) and (8). In the quasistatic limit of k_{0} L ≪ 1 the expression for the transfer matrix of the dielectric layer can be approximated as:(9)
Using this expression for A_{∞} and B_{∞}, we solve the system of equations (7) and (8). The solution of this system of equations is:(10) (11)where, ε_{A,eff} ≡ ε_{eff} is the effective epsilon of the whole fractal. Note that, in the quasistatic limit, the effective permittivity of the structure is expressed as a weighted average of the permittivities of the layers.
In order to understand the range of applicability of the obtained quasistatic expressions for the effective permittivity, we compare the eigenvalues of the total transfer matrices calculated by the effective medium approach and by the multilayer approach.
The transfer matrix eigenvalues for both the effective medium approach and the direct numerical calculation for multiple layers are shown in Figure 3. This figure depicts the real part and the absolute value of the imaginary part of the two eigenvalues as functions of k_{0}L. The curves representing these functions (the blue and golden curves) coincide within the propagation bands and split (form “loops” visible in Figure 3) in the stopbands. Figure 3 shows that outside of the stopbands and when k_{0}L ≲ 2, the result from the effective medium calculations is similar to the result from numerical calculations for 9th order.
Figure 3. The imaginary and real parts of eigenvalues obtained from numerical results and effective medium calculations for the 9th order fractal MTM. 
3 FDTD simulations
In order to verify the analytical results obtained with the transfer matrix approach, MIT PhotonicBands (MPB)^{1} package and MEEP^{2} simulation software package have been used [12].
In contrast with the transfermatrix based calculations, in which considering structures that have thousands of dielectric layers is feasible, in FDTDbased numerical calculations working with higher orders of fractals is more difficult. Therefore, the following results are given only for 9th and 11th order fractal MTMs which have 683 and 2,731 dielectric layers in the whole structure, respectively. Figure 4 shows the ε profile of the fractal of 9th order.
Figure 4. ε profile for the 9th order fractal MTM. Due to the limited resolution of the figure, the very thin dielectric layers located in the colored regions are not seen. 
3.1 FDTDbased band diagram calculations
Figure 5a depicts the permittivity profile of the 9th order fractal MTM as it is represented by the MPB. With increasing the resolution of the figure the distribution of ε is better seen for higher orders.
Figure 5. ε profile for the 9th and 11th order fractal MTM by MPB. The black and gray color are ε_{1} and ε_{2} respectively. (a) 9th order, (b) 11th order. 
The number of layers in each order of this structure is . The smallest layer thickness equals to Lr ^{ n } for , and L(1 − 2r)r^{n−1} for . Considering these conditions, the physical limits for the highest order of this structure can be estimated.
Figure 6 shows the band diagram for the 9th and 11th orders (the epsilon profile of this order is shown in Figure 5b) of fractal MTM. As is seen from the band diagram for the 9th order, the simulation results follow closely the analytical results shown in Figure 2.
Figure 6. Band diagrams for 9th and 11th order by MPB. 
As it is seen from Figure 6, since the finest fractal elements are already too small in these orders (as compared to wavelength), the band diagrams are very close to each other.
Figure 7 shows the agreement between analytical results and numerical calculations for the fractal of 9th order.
Figure 7. Comparison between band diagrams from analytical and numerical calculation for a 9th order fractal MTM. 
3.2 Field distributions in the propagation bands
Using the FDTD algorithm allows us also to study how the electromagnetic field of the modes propagating inside the structure is distributed within the multilayer structure.
The field distribution along the fractal structure is shown in Figure 8, for four bands of propagation. In these examples, . As one can see, the field distributions closely resemble the ones for plane waves propagating in uniform media, which shows that the studied fractal structure can be considered as effectively homogeneous in these frequency bands.
Figure 8. Distribution of the transverse electric field (real and imaginary parts) inside the fractal structure for different propagation bands (in units relative to the field maximum). 
4 Conclusion
In this work a new 1D metamaterial which is based on a fractal set has been presented. A homogenization method that uses selfsimilarity property of the structure has been developed. We have derived the effective material parameters of this fractal MTM structure and verified them numerically. The quasistatic effective permittivity obtained with this method is expressed as a weighted average of permittivity of dielectric layers in the fractal. This result agrees well with the known quasistatic solutions for layered media, which confirms the validity of the proposed homogenization method.
Acknowledgments
This work, is funded by FCT/MEC through national funds (PD/BD/52589/2014) and when applicable cofunded by FEDER PT2020 partnership agreement under the project UID/EEA/50008/2013.
References
 V.G. Veselago, Sov. Phys. Usp. 10 (1968) 509. [CrossRef] (In the text)
 J.V. Kranendonk, J. Sipe, Progress in Optics 15 (1977) 245–350, DOI: 10.1016/S00796638(08)704803. [CrossRef] (In the text)
 D.R. Smith, J.B. Pendry, J. Opt. Soc. Am. B 23 (3) (2006) 391, DOI: 10.1364/JOSAB.23.000391. [CrossRef]
 O. Ouchetto, C.W. Qiu, S. Zouhdi, L.W. Li, A. Razek, IEEE Trans. Microwave Theory Tech. 54 (2006) 3893, DOI: 10.1109/TMTT.2006.885082. [CrossRef]
 C.R. Simovski, Opt. Spectrosc. 107 (2009) 726, DOI: 10.1134/S0030400X09110101. [CrossRef]
 E. Poutrina, D. Huang, D.R. Smith, New J. Phys. 12 (2010) 093010. [CrossRef]
 A. Alù, Phys. Rev. B 84 (2011) 075153, DOI: 10.1103/PhysRevB.84.075153. [CrossRef] (In the text)
 K. Falconer, Fractal Geometry: Mathematical Foundations and Applications, 3rd edn., John Wiley & Sons, Inc, UK, 2014. (In the text)
 D.M. Pozar, Microwave Engineering, 4th edn., John Wiley & Sons, Inc, UK, 2011, https://books.google.pt/books?id=JegbAAAAQBAJ. (In the text)
 S. Tretyakov, Analytical Modeling in Applied Electromagnetics, Artech House, USA, 2002. (In the text)
 J.D. Joannopoulos, S.G. Johnson, J.N. Winn, R.D. Meade, Photonic Crystals, 2nd edn., Princeton University Press, USA, 2008, http://press.princeton.edu/titles/8696.html. (In the text)
 A.F. Oskooi, D. Roundy, M. Ibanescu, P. Bermel, J. Joannopoulos, S.G. Johnson, Comput. Phys. Commun. 181 (2010) 687, DOI: 10.1016/j.cpc.2009.11.008. [CrossRef] (In the text)
Cite this article as: Moeini S & Pedro JC: Extraction of material parameters in 1D fractal metamaterial. EPJ Appl. Metamat. 2016, 3, 7.
All Figures
Figure 1. ε profile for different orders of fractal MTM. In this example, L = d_{1} + 2 × d_{2} = 1, d_{2} =0.45L, ε_{1} = 4 and ε_{2} = 1. (a) 1st order, (b) 2nd order, (c) 3rd order, (d) 4th order. 

In the text 
Figure 2. The band diagram for the 9th order fractal MTM. Four propagation bands and three band gaps are seen in the diagram. 

In the text 
Figure 3. The imaginary and real parts of eigenvalues obtained from numerical results and effective medium calculations for the 9th order fractal MTM. 

In the text 
Figure 4. ε profile for the 9th order fractal MTM. Due to the limited resolution of the figure, the very thin dielectric layers located in the colored regions are not seen. 

In the text 
Figure 5. ε profile for the 9th and 11th order fractal MTM by MPB. The black and gray color are ε_{1} and ε_{2} respectively. (a) 9th order, (b) 11th order. 

In the text 
Figure 6. Band diagrams for 9th and 11th order by MPB. 

In the text 
Figure 7. Comparison between band diagrams from analytical and numerical calculation for a 9th order fractal MTM. 

In the text 
Figure 8. Distribution of the transverse electric field (real and imaginary parts) inside the fractal structure for different propagation bands (in units relative to the field maximum). 

In the text 