ISSN 1087 6596, Glass Physics and Chemistry, 2010, Vol. 36, No. 5, pp. 521553.

Pleiades Publishing, Ltd., 2010. Original Russian Text A.A. Zhilin, M.P. Shepilov, 2010, published in Fizika i Khimiya Stekla.

Metamaterials: A New Direction in Materials Science

AbstractA review is presented of materials science of metamaterials, i.e., the area of research that has been extensively developed over the last decade. Metamaterials are artificial materials consisting of structural ele ments, whose type and mutual arrangement can be specified during the fabrication. The development of metamaterials was associated initially with the idea of the design of electromagnetic media with a negative refractive index and, more recently, with the prospect for fabrication of superlenses, invisible objects, and other optical devices. As a result, there appeared a new branch of materials science and an intimately related new branch of opticstransformation optics. This review has discussed the main directions of research in this field and problems encountered in this way. It is noted that metamaterials cannot be designed without invoking the most modern tools for performing numerical simulation and that their implementation requires the use of high level technologies. The idea of metamaterials is important not only for optics and electromag netism but also for acoustics. Key words: metamaterials, negative refractive index, superlens, invisibility cloaks, transformation optics, nanostructuring DOI: 10.1134/S1087659610050019

INTRODUCTION In the last decade, metamaterialsartificial mate rials designed for the use in the filed of electromagne tism, including optical applicationshave been intensively developed all over the world. Metamateri als have been engineered from structural elements with the type and mutual arrangement that can be specified during the fabrication. If the size of structural elements in a metamaterial and the characteristic dis tance between them are considerably smaller than the radiation wavelength, the metamaterial can be consid ered as a continuous medium characterized by the permittivity and the permeability. In this case, the structural element of the metamaterial is an analog of an atom, a molecule, or a unit cell of a conventional material. Modern technologies, including nanotechnolo gies, and prospects of their development have opened up wide possibilities for producing metamaterials with a controlled (preliminarily calculated) structure that can exhibit unusual or even unique electromagnetic properties. It should be noted that the theory plays a key role in the design of these metamaterials: initially, theoretical principles have been used for choosing the type of structures that could ensure necessary electromag netic properties, then, quantitative characteristics of the structure have been selected and optimized using the simulation of the electromagnetic response, and,

thereafter, the material has been fabricated. The prep aration of metamaterials with controlled properties without preliminary theoretical calculations seems to be unrealistic. Important milestones on the way of the develop ment of this new direction in materials science were an idea of the creation of superlenses based on metama terials (2000) and the experimental preparation of the negative refractive index metamaterial for radiation with a frequency of 10.5 GHz (2001). The latter fact has demonstrated the possibility of engineering metamaterials with properties that are not observed for natural materials. At present, the works concerning an increase in the operating frequency of negative refractive index metamaterials have been continued. Progress in the fabrication of metamaterials has allowed one to raise the question as to the possibility of producing invisible objects. An intensive discussion of the invisibility problem, which was started in 2005, has been carried out in two directions. In the first approach to the solution of this problem, the case in point is objects with sizes that are smaller than or of the order of the electromagnetic radiation wavelength. The second approach considers the possibility of cre ating invisible cloaks that conceal the objects con tained in them from an external observer. The cloaks in the second approach should be completely prepared and the objects in the first approach should be partially fabricated from materials with unusual electromag

521

522

ZHILIN, SHEPILOV

ray in medium 1) and the angle of refraction (the angle between the normal and the refracted ray in medium 2) so that the incident and refracted rays lie on different sides of the normal to the interface [1]:
1 2

n sin = 2, sin n1

( ni =

i i, i = 1, 2 ).

(1)

Fig. 1. Refraction of light in passage to a conventional (right handed) material and to a material with a negative refractive index (left handed material) (according to [3]): (1) incident ray, (2) reflected ray, (3) refracted ray in the left handed medium, and (4) refracted ray in the conven tional medium. The horizontal line indicates the interface between the media.

netic properties and, generally speaking, can be pro duced only with the use of metamaterials. The second approach to the solution of the invisi bility problem is based on the transformation of coor dinates (mapping method). In recent two or three years, this method has been used for solving a number of problems of theoretical optics. We can argue that a new direction in optics, i.e., the so called transforma tion optics, has been formed to date. Devices that have been theoretically developed in the framework of this direction also cannot be implemented without use of metamaterials. This paper represents a review of the literature on the problem of metamaterials and related problems. It should be noted that, in a number of works, the terms refractive index, invisible, transparent, etc. have been used for electromagnetic radiation with a wider spectral range as compared to the visible light region, namely, for radiation with wavelengths from several centimeters (gigahertz frequencies) to several hundred nanometers (visible light). We will use the same extended version of the terms. NEGATIVE REFRACTIVE INDEX METAMATERIALS Hypothetical Left Handed Media (Media with Negative Refractive Index) A light ray passing from medium 1 to medium 2 changes its direction: it is refracted. The refraction law (Snells law) relates the angle of incidence (the angle between the normal to the interface and the incident

Here, ni is the refractive index of the ith medium, which is expressed in terms of the permittivity i and permeability i (as a rule, it is assumed that i = 1 in the optical range of frequencies [2]; however, for the purpose of further presentation, we will consider the general case i 1). Figure 1 taken from [3] shows the refraction of the ray at the interface between two media. The incident and refracted rays are designated as 1 and 4, respectively, and it is assumed that the media are transparent and that medium 2 is optically denser; i.e., n2 > n1 1. Mandelstam [4] was first to consider and Veselago [3] thoroughly analyzed the case of the so called neg ative refraction. Veselago examined the propagation of electromagnetic waves in a hypothetical medium with negative values of the permittivity < 0 and permeabil ity < 0. He noted that, for a plane monochromatic wave propagating in this medium, the electric field vector E, the magnetic field vector H, and the wave vector k form a left handed vector triple, whereas these vectors in conventional (right handed) media form a right handed vector triple. These hypothetical media were referred to by Veselago as the left handed media. In the left handed medium, the energy flux density vector (Poynting vector) and the group velocity vector (coinciding with the former vector in direction) are antiparallel to the wave vector k and, hence, to the phase velocity vector [3]. It should be noted that con sideration of a negative phase velocity goes back to the works by Lamb [5] (in hydrodynamics) and Schuster [6] (in optics), as well as to the more detailed work by Mandelstam [7] (solid state physics). Veselago [3] showed that the Snells law (1) holds true for any combinations of right handed and left handed media in the use of the generalized definition of the refractive index: n1 = pi i i , p i = 1 ( i > 0, i > 0 ), p i = 1 ( i < 0, i < 0 ); (2)

i.e., under the assumption that the refractive indices of left handed media are negative. It should be noted that, in media with < 0 and > 0 or > 0 and < 0, waves attenuate rapidly [8]. In passage from the conventional (right handed) medium with the refractive index n1 1 to the left handed medium with the refractive index n2 < 0, the incident and refracted rays lie on the same side of the normal to the interface (Fig. 1, rays 1, 3) in contrast to
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

523

the case of conventional refraction (rays 1, 4). In the construction of Fig. 1, it was assumed that |n2| > n1. Veselago [3] noted that, if light passes from the medium with 1 > 0 and 1 > 0 to the medium with 2 = 1 and 2 = 1, the reflected ray is absent. The case where the first medium is vacuum (1 = 1 and 1 = 1) and 2 = 2 = 1, i.e., the refractive index of the sec ond medium is n = 1, is most interesting. This case is associated with one of the possible applications of media with a negative refractive index. Simple geo metric considerations demonstrate [3] that, in vac uum, a plate from a left handed material with a thick ness d and refractive index n = 1 focuses radiation of a point source located at a distance l < d from the plate at a point. In other words, this plate is a lens in which the image formation is clearly illustrated in Fig. 2 taken from [9]. As was theoretically showed by Pendry [10], this lens is perfect (superlens) in the sense that its use makes it possible to overcome the diffraction resolution limit associated with the wave nature of light. The Pendry idea of the superlens, in addition to the prediction of the structure and fabrication of neg ative refractive index metamaterials (see below), resulted in a sharp intensification of investigations in this field. However, the superlens concept came under criticism [1113]. The answers to the questions for mulated in [1113] were given in [1417], where the superlens concept and its limitations were refined. The superlens was numerically simulated in [18, 19]. Moreover, the problem regarding the superlens was also discussed in [8, 9, 20, 21]. The possibility of over coming the diffraction resolution limit with the use of a plane parallel plate with a negative refractive index (superlens) was first experimentally proved for photo nic crystals in [22] and metamaterials in [23]. Finally, reasoning from the requirement of a pos itive field energy, Veselago [3] made the inference that, first, the values of and can be simultaneously negative only in the presence of a frequency disper sion and, second, the frequency dependences of the permittivity and permeability should satisfy the con ditions ( ) > 0, ( ) > 0, (3)

Fig. 2. Image formation by a plane parallel plate produced from a material with the refractive index n = 1 [9].

history of the appearance of ideas associated with the left handed media. Development of Samples of Negative Refractive Index Metamaterials The permittivity and permeability characterize a macroscopic response to applied electric and mag netic fields. These are macroscopic parameters that are determined in the course of macroscopic measure ments as a result of the averaging over fairly large time intervals and spatial volumes. For conventional mate rials, the characteristic scale of the spatial averaging should considerably exceed the scale of the material inhomogeneity, which is determined by the size of the volume per atom (molecule) or the unit cell size in the case of crystalline solids. Now, we explain why it is impossible to produce conventional materials with a negative refractive index in the near IR and visible spectral ranges [8]. The appearance of negative values of the permittivity and permeability is associated with the resonance phe nomena in the response of the medium to the electro magnetic action. As in the case of any resonance, the response follows in phase with the applied field at fre quencies below the resonant frequency. In some range of frequencies exceeding the resonant frequency, the response follows in antiphase with the applied field, which provides the possibility of observing the negative permittivity or permeability if the resonance is suffi ciently sharp. There are conventional materials char acterized by a strong electric response at almost any specified frequency in the range from radio frequen cies to frequencies of UV radiation. However, the magnetic response in the majority of conventional materials is limited by low frequencies of the micro wave range. Therefore, the permeability for these
2010

where is the angular frequency. Here, it is pertinent to note one more limitation. As was noted in [8], there are no fundamental objections against the negativity of the real parts of the quantities and . At the same time, the absorbed energy in a thermodynamically equilibrium medium should be positively definite, which leads to negative values of the imaginary parts of the quantities and . In a ther modynamically nonequilibrium medium (gain medium), this requirement fails. In conclusion of this section, we note the reviews [24, 25] that contain interesting information on the
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

524

ZHILIN, SHEPILOV

Fig. 3. Two dimensional system of metal cylinders [43, 44] for which the frequency of the longitudinal plasmon mode (expression (4)) is determined by the geometric parame ters of the structure.

materials almost does not differ from unity already at frequencies substantially lower than the optical fre quencies [2, p. 374]. The magnetic polarization, as a rule, is caused either by unpaired electron spins or orbital electron currents, and collective excitations of this type usually occur at low frequencies [8]. Some ferromagnetic, ferrimagnetic, and antiferromagnetic materials exhibit a specific magnetic activity at fre quencies up to several hundred gigahertz [2631]. However, these materials are rare and have narrow bands of the magnetic activity. At the same time, the materials artificially structured on microscales and nanoscales, i.e., metamaterials, can have a magnetic activity at almost any frequencies in the range from several hertz [32] to frequencies of the near IR range [33], including radio frequencies [34, 35], microwave frequencies [36, 37], and frequencies from several units [38] to several hundred terahertz [39]. In regard to optical magnetism, it was noted in [40] that, at present, its implementation [41, 42] is accompanied by a high level of losses. In this respect, artificial materials of the new type, i.e., metamaterials, have been treated in the last decade as promising materials for the observation of the negative refractive index; in this case, the response to the electromagnetic field is determined by the behavior of structural elements forming the metama terial and the size of the structural elements signifi cantly exceeds atomic sizes. In order for the notions of the permittivity and permeability can be applied to metamaterials, it is necessary to assume that the size of structural elements of the metamaterial should be considerably smaller than the radiation wavelength. It is these materials characterized by a negative refractive

index that below will be called the negative refractive index metamaterials in contrast to photonic crystal line structures in which there can occurs a negative refraction but the radiation wavelength is of the order of the lattice constant of the photonic crystal (i.e., the notions of the permittivity and permeability in a con ventional sense are inapplicable to consideration of negative refraction in photonic crystals) [8]. Extensive investigations of metamaterials traces back to the Pendry ideas of structured artificial mate rials that can have a negative permittivity [43, 44] or a negative permeability [36]. The permittivity of metals is determined by the response of a free electron gas that is characteristic of a plasma and is negative in the UV spectral range [8]. These materials can be characterized by numerous resonance states localized on their surface and referred to as surface plasmons [45, 46]. An analysis demon strates [8, 46] that surface plasmons cannot be excited on a perfect flat surface with the use of propagating light modes; i.e., in order to excite these plasmons, it is necessary to provide coupling mechanisms, such as the surface roughness, lattice structures, or dielectric con nection to vacuum (hemisphere, prism). The interac tion of surface plasmons with radiation on structured surfaces offers new possibilities in plasmonics. In order to obtain a negative permittivity in the gigahertz spectral range, Pendry et al. [43, 44] pro posed to use two dimensional and three dimensional structures consisting of cylindrical metal wires. In the three dimensional structure, the wires form edges of a simple cubic lattice. In the two dimensional structure, the parallel wires are arranged so that their cross sec tions by a perpendicular plane form a square lattice (Fig. 3). The frequency dependences of the permittiv ity were analytically calculated and numerically simu lated [43, 44] and measured experimentally [44] for these structures. The corresponding results allowed one to make the inference that the frequency of plas mons in these systems lies in the gigahertz range for a wire diameter of the order of a micron and a charac teristic interwire distance of the order of several milli meters. For example, for the two dimensional system shown in Fig. 3, the analytical expression for the fre quency of the longitudinal plasmon mode p (the electric field vector is directed along the wires) has the form p =
2

2c , 2 a ln ( a/r )

(4)

where r is the wire diameter, a is the constant of the two dimensional lattice (see Fig. 3), and c is the speed of light in free space. For r = 1 m and a = 5 mm, relation ship (5) gives p = 8.2 GHz, which is in good agreement with the frequency p = 8.3 GHz determined by the numerical simulation [44]. The wavelength of this radi ation is p 4 cm, and the condition for the applicability
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

525

Fig. 4. Two possible variants of the split ring resonator (see text for explanation).

of the effective medium approximation p a is ful filled. The example of the experimental implementa tion of the system of this type at higher frequencies is described in [47]. In this case, r = 15 m and a = 120 m, and the frequency p = 0.7 THz calculated from relationship (4) agrees well with the observed fre quency and the condition for the applicability of the effective medium approximation is also satisfied (p/a 3.6). Therefore, a change in the geometric parameters of this structured system makes it possible to control the position of the spectral range of the negative permittivity. A negative permeability in the gigahertz frequency range can be obtained in a structured system that is built up from nonmagnetic conductors, i.e., resona tors each of which consists of two planar split rings [36]. Two possible variants of the split ring resonator are depicted in Fig. 4. The alternating magnetic field directed perpendicular to the ring plane excites the electric current in the rings, which are inductance capacitance elements: the rings themselves are induc tances and the gap between the rings and cuts in the rings form capacitors. The current excitation has a res onant character, and the resonant frequency of this LC resonator depends on the geometric parameters. Initially, Pendry et al. [36] considered a system of identical cylindrical resonators with parallel axes. The points of intersection of the resonator axes with the plane perpendicular to these axes form a square lattice with the period a. Each resonator consists of two con ducting coaxial cylindrical layers that have different diameters and are cut along the lines parallel to the axis so that the cross section of the resonator by the plane is a split ring resonator shown in Fig. 4a. An approximate formula was derived for the permeability of this structured system when the magnetic field is oriented along the resonator axes under the assump tion that the gap between the rings d (Fig. 4a) is small
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

as compared to the ring radius r. The effective perme ability has the resonance form eff ( ) = 1 + where f = r /a ,
2 2

f , 2 2 0 i

(5)

3d 1/2 0 = , 2 r 3 0 0

2 . (6) 0 r

Here, a is the period in the arrangement of resonators; 0 and 0 are the permittivity and permeability of free space, respectively; is the relative permittivity of the compound in the gap between rings (in [36], = 1); and is the resistance of the cylindrical resonator per unit square of its surface. The factor f is the filling fac tor of the material. The dependence eff() (relationship (5)) has a resonant character with the resonant frequency 0; in this case, the resonance is associated with the induc tancecapacitance resonance in the system. For a small attenuation (i.e., in the case of a low resistivity when 0), the response does not coincide in phase with the applied magnetic field at frequencies > 0 and the quantity eff is negative in the range m > > 0, where m is the magnetic plasma frequency, which, by disregarding the resistivity of the material, can be rep resented in the form [36] 3d 1/2 . m = ( 1 f ) 2 r 3 0 0 (7)

Formula (7) indicates that the filling factor plays a decisive role in the width of the range m0 in which eff () < 0. The inclusion of the resistivity of the material broadens the resonance, and, in the case of the material with a high resistivity, the resonance is suppressed so that the range of negative values of the real part of the permeability eff() disappears [8].
2010

526 (a) 5 Re(eff)

ZHILIN, SHEPILOV

1 2 3

5 2 3 (b) Im(eff) 10 1 2 3 4

Fig. 6. Unit cell of the isotropic metamaterial fabricated from split ring resonators [36].

0 2 3 4 Frequency , GHz

Fig. 5. Frequency dependences of the (a) real and (b) imag inary parts of the permeability eff() (relationship (5)). The calculations were performed for r = 2 mm, d = 0.1 mm, a = 5 mm, = 1, and three values of the resistivity = (1) 0.2, (2) 1.0, and (3) 7.0 .

This statement is illustrated by Fig. 5. This figure was constructed with the geometric parameters used for the structure in [36, relationships (26)]. For small val ues of ( 1 ), the system is characterized by the negative real part of the permeability in the frequency range 2.95 GHz < = /(2) < 4.16 GHz with the limits determined by the values of 0 (expression (6)) and m (expression (7)). This range narrows with an increase in the resistivity in the range 1 < < 6.8 . Finally, at > 6.8 , the real part of the permeability is positive for any frequencies ( = 7 for curve 3 in Fig. 5). It should also be noted that, in the frequency range < 4.5 GHz (shown in Fig. 5), the radiation wavelength is > 6.6 cm; i.e., it is larger than the period a = 5 mm by a factor of more than ten, and, hence, the condition for the effective medium approximation is satisfied. This system of cylindrical resonators does not exhibit a magnetic activity when the magnetic field is directed perpendicular to the axes of the cylinders; i.e., the metamaterial is uniaxial. In order to overcome this limitation, Pendry et al. [36] proposed to replace the cylindrical resonator under consideration by a set of planar ring resonators (Fig. 4). The centers of the ring resonators are equidistantly arranged on the axis with the period l, their planes are perpendicular to the axis, and the ring orientation is identical and corre sponds to the orientation of the cross sections of the

cylindrical resonator. By replacing the cylindrical res onators in the above system by sets of ring resonators and assuming that l < r, the authors derived the rela tionship for the permeability of this system in the direction of the axes and demonstrated that the per meability can have negative values. The authors also noted that the continuous electrical path that is characteristic of cylindrical resonators is absent in the system of ring resonators. Like the initial system of cylindrical resonators, the obtained system of ring res onators is uniaxial; however, the use of ring resonators allows one to construct the isotropic metamaterial, for which the unit cell is shown in Fig. 6 [36]. It should be noted that the equality l = a is valid for this metamate rial and the inequality l < r used to derive the formula for the permeability does not hold true, because the ring diameter should be smaller than the lattice spac ing; i.e., r < a/2. Other inequalities used for deriving the formula can also be invalid in real systems. How ever, the authors believe that the nonfulfillment of the inequalities has an effect only on the accuracy of the derived relationships rather than on the functional pos sibilities of the structures under consideration. Addi tional details of the electromagnetic response of split ring resonators were theoretically investigated in [48]. The use of split ring resonators made it possible to fabricate metamaterials with the magnetic activity at frequencies of approximately 1 GHz [49], 5 GHz [37], 10 GHz [5052], 100 GHz [53], 1 THz [38], and 30 THz [54]. The magnetic resonance at higher fre quencies of 100 THz [39] was observed for systems of single rectangular rings (in contrast to double rings shown in Fig. 4) with sizes of 320 nm. For a similar system containing rings with a smaller size (200 nm), the magnetic resonance was observed at frequencies of 200 THz ( = 1.5 m) in the case where the light
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

527

E k H
Fig. 7. Metamaterial [37] for which the permittivity and permeability are negative for radiation with a frequency of 5 GHz, the electric vector parallel to wires, the magnetic vector perpendicular to the planes of ring resonators, and the wave vector parallel to the ring plane and perpendicular to wires. Shown is the right handed vector triple (E, H, k) for the incident wave (waves in empty space). In the medium, the vectors have opposite directions and form the left handed triple. The lattice constant is a = 8 mm, the wire radius is 0.8 mm, the inner radius of the inner ring is r = 1.5 mm, the ring width is w = 0.8 mm, and the distance between the rings is d = 0.2 mm (see designations in Fig. 4a).

wave vector was directed along the normal to the ring plane and at frequencies of 370 THz ( = 800 nm) in the case of oblique incidence of light [55]. The mini mum permeability was observed at the wavelength = 1.67 nm and was = 0.25. It should also be noted that the structures characterized by the magnetic activity in the radio frequency range were fabricated using more complex elements, i.e., the so called Swiss rolls [35], which were previously studied theoretically in [36]. The first metamaterials with negative values of the permittivity and permeability for radiation with a spe cific frequency, i.e., negative refractive index metama terials, were prepared by combining the structure of parallel wires (Fig. 3) with periodic arrays of split ring resonators (Fig. 4). The simplest metamaterial of this type (Fig. 7) was proposed and experimentally investi gated by Smith et al. [37]. The authors made the infer ence that the metamaterial is characterized by nega tive values of the permittivity and permeability with respect to radiation with a specific frequency and polarization (see the caption of Fig. 7). However, in view of its structure this metamaterial is substantially anisotropic. More recently, Shelby et al. [50] proposed the metamaterial that was isotropic in two dimensions and consisted of copper split ring resonators and copper wires (strips) (Fig. 8). The unit cell size a was equal to 5 mm. The array of parallel wires (vertical wires in Fig. 8) provides a negative permittivity in the gigahertz frequency range in the case where the electric field is directed along the wires. The array of split ring reso nators ensures a negative magnetic permeability in the case where the magnetic field vector lies in the plane
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

perpendicular to the wires (the horizontal plane in Fig. 8), so that the permeability in this plane does not depend on the direction of the magnetic field vector in the framework of the effective medium approxima tion (two dimensional isotropy of the system). Conse quently, this metamaterial should have a negative refractive index for electromagnetic radiation with the wave vector oriented perpendicular to the wires (or, to put it differently, radiation polarized parallel to the wires), and, for this orientation, the refractive index is two dimensional isotropic. By using the prism from this metamaterial, Shelby et al. [50] thoroughly inves tigated the refraction of radiation at the frequency = 10.5 GHz (the wavelength 3 cm is six times larger than the lattice constant a) and with the wave vector oriented perpendicular to the wire direction and led to the conclusion that the Snells law (1) holds true and the refractive index of the metamaterial is n = 2.7 0.1. (8) Reasoning from the results of the experimental inves tigations of the frequency dependence of the refrac tive index [50, Fig. 4], the authors assumed that the refractive index is negative in the frequency range from approximately 10.2 to 10.8 GHz; i.e., in this range, both the permittivity and the permeability are negative. The metamaterial studied in [51] differs from the metamaterial investigated in [37] (see Fig. 7) by the presence of the additional wire (in the unit cell) with the same direction as that shown in Fig. 7, the square shape of rings, and sizes of elements. In [51], the experimental investigation and calculations were per
2010

528 (a)

ZHILIN, SHEPILOV (b) z

Fig. 8. (a) Two dimensional metamaterial [50] characterized by the negative refractive index n = 2.7 0.1 at a frequency of 10.5 GHz and (b) its fragment. The lattice constant a is 5 mm.

formed for radiation with a specified orientation of the wave vector and the electric and magnetic field vec tors, which is shown in Fig. 7 (correspondingly, the authors characterize the metamaterial as one dimen sional). The refraction experiments showed that, for the radiation with the above orientation, the refractive index in the frequency range 12.613.2 GHz is nega tive and the real part of the refractive index at a fre quency of 12.6 GHz is equal to 1.05. The experimen tal values of the refractive index are in good agreement with the values obtained by numerically solving the Maxwells equations. It should be noted that the infer ence made in [51] regarding the negative refractive index is based on the direct experiments on the refrac tion, whereas Smith et al. [37] used the indirect data on the transmission. Gokkavas et al. [53] fabricated the metamaterial of the same type but with three wires in the unit cell with the refractive index n 5 in the frequency range 98104 GHz. Planar structures in which the unit cells have a size of approximately 100 m and consist of a split ring resonator (Fig. 4a) and a wire piece were prepared by Moser et al. [56]. The authors studied the transmission spectra and made the inference that these structures are electromagnetic metamaterials for radiation with a frequency of the order of 1 THz. The properties of these structures were not investigated in more detail. We do not know works in which double or single split ring resonators were used for fabricating negative refractive index metamaterials at frequencies higher than those achieved in [53]. However, there exists one more approach to the design of metamaterials characterized by a magnetic activity with the use of localized plasmon resonances without classical split ring resonators [57]. Let us con sider this approach. Lagarkov and Sarychev [58] noted that a pair of noble metal nanorods can exhibit a large paramagnetic response, and Podolskiy et al. [59] the

oretically demonstrated that this pair can also be char acterized by a magnetic response at a wavelength of 1.5 m. The fact that a set of these pairs can have a negative permeability at a wavelength of 1.5 m was experimentally confirmed in [60]. In [59], it was pre dicted for the first time that this pair of rods can have a negative value of the real part of the refractive index in the visible spectral range, and the problem under con sideration was discussed in more detail in [6163]. For example, the calculations showed that a pair of parallel cylindrical gold rods 162 nm in length, 32 nm in diam eter, and with a distance of 80 nm between the axes has negative electric and magnetic responses in the wave length range 500600 nm in the case where the elec tric vector of the incident plane wave is directed along the rods and the magnetic vector is perpendicular to the plane of the rods. The qualitative aspects of the appearance of resonances in this system were dis cussed by Klar et al. [57]. The system consisting of pairs of gold nanorods was experimentally investigated by Shalaev et al. [64]. The two dimensional regular array of pairs of nanorods on a glass substrate was prepared using electron beam lithography, so that the plane containing a pair of rods was perpendicular to the substrate. The sample, its preparation, and the technique for measuring the refractive index are described in detail in [42, 65]. In [64], the inference was made that, for light that is inci dent perpendicular to the substrate and polarized identically to that described in the above calculations [63] (see the previous paragraph), the refractive index of the metamaterial is negative in the wavelength range 1.41.6 m. At a telecommunication wavelength of 1.5 m, the refractive index was equal to 0.3 0.1. The results of the measurements of the spectral depen dence of the refractive index are in agreement with the results of the performed numerical simulation.
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

529

However, here, we should note the following. In [64], the refractive index was determined from the data on the complex reflection and transmission coeffi cients of a thin metamaterial layer (i.e., by the indirect method [66]) in the case where the light flux was per pendicular to the substrate on which pairs of nanorods were located. Therefore, the radiation passes through the metamaterial that is not a bulk layer. This layer can be treated as monomolecular because a pair of nan orods is a molecule of the metamaterial. In this case, the losses, which can be evaluated from the experimental data on the transmission and reflection [64, Fig. 2a] are no less than 15% in the spectral range 1.41.6 m, in which the real part of the refractive index is negative. For example, the losses at a wave length of 1.5 m are more than 30%. This means that the bulk material under consideration would be opaque and the possibility of directly measuring its refractive index seems to be problematic. This corre sponds to large values of the imaginary part of the complex refractive index; for example, n 0.3 + i2.7 at a wavelength of 1.5 m [64, Fig. 3b]. Grigorenko et al. [67] fabricated systems consisting of pairs of gold nanodots (objects with sizes of approximately 100 nm) on a substrate and investigated their interaction with polarized light incident perpen dicular to the substrate. Systems of this type also rep resent a monomolecular layer of the metamaterial. The authors established that these systems can have negative values of the real part of the permeability in the visible spectral range. Moreover, they can have simultaneously negative values of the real parts of the permittivity and permeability; for example, in one of the cases under consideration, these values were esti mated to be Re() 0.7 and Re() 0.3 at some wavelength in the green range [67, p. 20]. Nonethe less, the authors noted that the negative refraction cannot be observed as a result of the large values of the imaginary parts (Im() 1i). By using the results of the numerical simulation, Zhang et al. [68] predicted and prepared [69, 70] one more type of metamaterials with a negative refractive index in the IR spectral range. The authors considered a three layer planar structure (Fig. 9) with a system of holes that form a two dimensional lattice (fishnet structure). The central layer is a dielectric (Al2O3) coated by gold films on both sides. The thickness of the dielectric layer was equal to 60 [68, 69] and 75 nm [70], and the thickness of each coating was 30 nm. The sizes and the period in the arrangement of holes were of the order of several hundred nanometers. The calcula tions [68] carried out for four structures with rectangu lar holes and characteristic sizes ax = ay = 801 nm, dx = 500 nm, and dy = 100, 200, 300, and 500 nm (see Fig. 9) demonstrated that these structures are charac terized by the negative refractive index for the radia tion with a wavelength of approximately 2 m, which is incident perpendicular to the layer and has a polar ization shown in Fig. 9. According to the calculations,
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

Fig. 9. Images of a fragment of the metamaterial proposed in [68]: (a) spatial image and (b) top view.

the figure of merit that is defined as the ratio (taken with the opposite sign) between the real (negative) and imaginary parts of the refractive index of the metamaterial = Re(n)/Im(n) (9)

can reach values of 26 for the structures under con sideration, which significantly exceed 0.1 obtained in [64] (see above). For example, the maximum value of = 6 was observed at 1.9 m for two structures with sizes dy = 100 and 200 nm. The experimentally fabricated three layer fishnet structure with circular holes [69] has a refractive index n 2.5 + i4 ( 0.6) for the radiation with a wave length of approximately 2 m, which is incident nor mal to the layer and has a particular polarization. As was noted in [70], this structure is characterized by a low transmission (<4%) not only due to the substantial absorption, which leads to a low value of , but also due to the mismatch between the impedances of the metamaterial and the environment (glass and air). The preparation of structures with elliptical holes of the same orientation [70] allowed one to decrease the mis match between the impedances and to increase the transmission to 40% and more in the spectral range of the negative refractive index ( 1.61.8 m). How ever, the level of losses remained considerable. It should be noted that, in [69], as in [64], the refractive index was determined by an indirect method, whereas this quantity in [70] was estimated by calculations. As in [64], the samples in [6870] represented a mono molecular layer of the metamaterial; i.e., they were not bulk systems. Klar et al. [57] noted that the structures
2010

530

ZHILIN, SHEPILOV

Fig. 10. Images of a fragment of the metamaterial in the form of a stack of three layer fishnet structures [75].

studied in [64] and [6870] are inverse with respect to each other: instead of pairs of metal rods in one struc ture [64], pairs of holes in two metal films are used in other structures [6870], which illustrates the use of the Babinet principle in the development of metama terials [71]. Dolling et al. [72] fabricated the metamaterial with the geometric structure similar to the structure of the metamaterial theoretically considered in [68] (see Fig. 9). The central layer was prepared from the MgF2 dielectric with the thickness s = 30 nm, the coatings with the thickness t = 45 nm were produced from sil ver, and the system of holes was characterized by the sizes ax = ay = 600 nm, dx = 316 nm, and dy = 100 nm. In [72], the authors noted that the attenuation for sil ver thin films in a wavelength range of 1.5 m is four times lower than that for gold films. The prepared fishnet material is characterized by a negative refractive index and a low level of losses in the spectral range 1.35 1.45 m; in particular, the maximum figure of merit 3 (relationship (9)) is observed at a wavelength of approximately 1.4 m for Re(n) 1. A decrease in the characteristic sizes of the structure of this type to s = 17 nm, t = 40 nm, ax = ay = 300 nm, dx = 102 nm, and dy = 68 nm [73] made it possible to decrease the operating wavelength of the metamaterial to 780 nm (Re(n) 0.6); however, the loss level in this case was more significant ( 0.5). Finally, Dolling et al. [74] experimentally demonstrated that a similar metama terial can be simultaneously characterized by negative values of group and phase velocities. Zhang et al. [75] calculated the electromagnetic properties of bulk metamaterials in the form of stacks of three layer fishnet structures for polarized radiation with a specified orientation (Fig. 10) with the wave

length in the range 1.52.2 m. The thickness of the air gap between structures was equal to 5 nm, and the parameters of the three layer structures (the dielectric thickness is equal to 60 nm, the thickness of gold films amounts to 30 nm, ax = ay = 801 nm, dx = 500 nm, dy = 200 nm) (see Fig. 9) were chosen to be identical to those in the optimum variant of the three layer structure considered in [68]. In [75], the calculations were performed for the stacks consisting of 1, 2, 5, 6, 100, and 200 three layer fishnet structures. The unit cell size for this metamaterial in the direction of prop agation of radiation (the thickness of the three dimen sional structure plus the thickness of the air gap) is equal to 125 nm; i.e., this size is substantially smaller than the radiation wavelength under consideration, whereas the thickness of the stacks consisting of 100 and 200 structures considerably exceeds the wave length. The calculations showed that, for one three layer fishnet structure, the real part of the refractive index is negative in the wavelength range 1.772.18 m. The changeover to the stack consisting of two and more structures leads to a shift in the range of negative values toward the short wavelength range (1.52.0 m). In this case, the results of calculations rapidly converge with respect to the number of structures in the stack, so that the calculated values of the real and imaginary parts of the refractive index are almost identical for the stacks containing five and ten structures. The imagi nary parts of the refractive index are small (smaller than 0.1 in the range 1.52.0 m), whereas the real parts of the refractive index noticeably exceed in mag nitude the values obtained for the single structure. As a result, the figure of merit of the stack (i.e., the bulk material) is considerably larger than that of the single structure [68] and reaches values of 25 in the wave length range of approximately 1.7 m. Therefore, Zhang et al. [75] theoretically showed that the stacks of fishnet structures are promising metamaterials for the observation of the negative refractive index in the IR spectral range. Moreover, Zhang et al. [75] also noted that the removal of air gaps between the structures in the stack depicted in Fig. 10 does not lead to a substantial change in the result of the calculations. In this case, the metamaterial is a layered structure in which metal and dielectric layers alternate with each other and there is a regular system of holes with axes perpendic ular to the layer planes. The bulk material of this type (Fig. 11) with the negative refractive index was described in [76]. The authors prepared the layered structure in which silver layers 30 nm thick alternated with magnesium fluoride (MgF2) layers 50 nm thick. The initial sample consisted of 21 layers. Focused ion etching was used to produce a regular system of rect angular holes with the period ax = ay p = 860 nm and sizes dx a = 565 nm and dy b = 265 nm. It should be noted that, in view of the specificity of the procedure for producing holes, their size decreased with an increase in the depth of the material (Fig. 11b). The
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE (a) (a)

531

a p

Ag MgF2 (b) 5 m 5 m (b)

n=1 f2

n<0

Prism 1 m
Fig. 11. Metamaterial in the form of a layered structure in which silver layers (light) alternate with magnesium fluo ride (MgF2) layers (dark) with a system of holes [76]: (a) schematic drawing and (b) scanning electron micro scope image of the real structure.

Quartz

Fig. 12. (a) Prism prepared from the layered metamaterial and (b) scheme of experiments on the study of refraction of IR radiation (according to [76]).

sample was used to prepare the prism by treating one of its faces, and the angle between the untreated and treated faces (the base angle of the prism) is designated as (Fig. 12). Two prisms with the angles = 5.0 and 4.7 were prepared. Light was incident perpendicular to the untreated face (Fig. 12b) and does not change its direction in the metamaterial, i.e., propagated accord ing to the scheme shown in Fig. 10. In this case, the angle of incidence on the treated face of the prism inside the metamaterial was equal to the angle . The measurement of the angle of refraction for the escape from the metamaterial (Fig. 12b) allowed one to calculate the refractive index of the metamaterial directly from the Snells law (1). If the refractive index of the metamaterial is negative, the refracted ray is located at the right of the normal to the surface shown by the dashed line in Fig. 12b. The measurements per formed in the spectral range 1.2001.775 m demon strated that, in this range, the real part of the refractive index monotonically decreases with an increase in the
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

wavelength from n = 0.63 0.05 at = 1.2 m to n = 1.23 0.34 at = 1.775 m; in this case, n < 0 for wavelengths from 1.540 to 1.775 m (the range with longer wavelengths was not studied in the work). In this case, the authors noted that the period of the metamaterial in the direction of radiation propagation (the sum of the thicknesses of the silver and magne sium fluoride layers) is equal to 80 nm, i.e., consider ably smaller than the radiation wavelength. The results of the experimental determination of the real part of the refractive index are in satisfactory agreement with the results of the simulation. However, the experimen tal and theoretical evaluations of the figure of merit differ substantially. At > 1.54 m, the experimentally determined values of increase monotonically with an increase in the wavelength and the maximum value of 3.5 is achieved for the longest wavelength = 1.775 m under investigation. The calculated depen dence () exhibits a maximum 18 at 1.72 m, and, at = 1.775 m, the simulation leads to 17.
2010

532

ZHILIN, SHEPILOV

The authors explained this substantial discrepancy between the experimental and theoretical values of by the imperfection of the prepared metamaterial. However, they noted that the obtained value of 3.5 is among the best experimentally determined values and express their hope that an improvement of the procedure for fabricating the metamaterial will allow one to approach the theoretical value of 20. It should be noted that, not long before the pub lishing of the work by Valentine et al. [76], the prism made from the metamaterial that represents a stack of fishnet structures was used by Navarro Cia et al. [77] to experimentally demonstrate the negative refractive index for the radiation at a frequency of approximately 55 GHz ( 5.5 mm). Fifteen metal plates 0.5 mm thick with a regular system of circular holes 2.5 mm in diameter were arranged in a stack not densely but with air gaps of 1 mm, which can be considered an analog of dielectric interlayers investigated in [76]. The system of holes in the plate [77] represented a rectangular planar lattice with the periods ax = 3 mm and ay = 5 mm. The prism prepared from this metamaterial with the angle = 26.6 was used in the scheme for measuring the refractive index, which is similar to the scheme depicted in Fig. 12b. The results of the measurements agree well with the results of the calculations. It was shown that the refractive index is close to zero at the frequency f 58 GHz, decreases with a decrease in the frequency, and is equal to 1 at f 53.5 GHz. There fore, the authors in [77] presented the bulk metamate rial with the negative refractive index; however, like the other bulk metamaterial of this type [76], it is anisotro pic and characterized by the obtained refractive index only for a specific direction of propagation and a par ticular polarization of radiation. The negative refractive index metamaterials in the visible spectral range were prepared by Lezec et al. [78]. The authors designed and fabricated the planar metamaterial based on nanosized metaldielectric waveguide structures. They directly observed the iso tropic negative refraction in two dimensions over a wide range 50 nm in the vicinity of the wave length 500 nm. For example, the produced mate rial at the wavelength = 514 nm was characterized by the refractive index n 5. The figure of merit was estimated to be high ( 4). The authors suggested that blocks from this metamaterial can be used for designing metamaterials characterized by negative refraction in three dimensions. As far as we know, the problem of fabrication of bulk isotropic metamaterials with a negative refractive index has remained unsolved to date. Molecules of this metamaterial should form a macroscopically homogeneous and isotropic system, for example, be located at sites of a three dimensional cubic lattice. Systems of this type that consist of fourfold split con ducting rings and crossing wires and are characterized by a negative refractive index were theoretically con

sidered in [79]. More complex metal structures with cubic symmetry that are intended for observing a neg ative refractive index in the IR spectral range were cal culated in [80]. Finally, as was proposed in [78] and noted above, bulk isotropic metamaterials with a neg ative refractive index can be prepared from blocks of two dimensional metamaterials based on waveguide structures. However, the practical implementation of the aforementioned and similar structures is a com plex problem, especially in the nanometer range of sizes. From our point of view, the simplest way of fabri cating bulk isotropic metamaterials with a negative refractive index consists in producing two phase sys tems with a random or regular distribution of phases, where one of the phases provides a negative electric response and the other phase ensures a negative mag netic response in a specified frequency range [81, 82]. Kussow et al. [81] theoretically investigated the metamaterial that consists of monodisperse spherical particles randomly arranged in a matrix (or regularly located at sites of a simple cubic lattice). The authors treated the polycrystalline magnesium diboride MgB2 as one of the phases (matrix) and the polaritonic mate rial SiC as the second phase (particles). The interac tion of the system of particles with electromagnetic radiation, which is described in terms of the Mie the ory [83] and the effective medium approximation in a particular form (see, for example, [81, 84]), has a res onant character, and there exists a frequency range in which the effective permeability of the metamaterial is negative. As was demonstrated in [81], the plasmon response of the matrix provides a negative permittivity in another frequency range. By properly choosing the volume fraction (f 0.3) and the particle radius (r 100 nm), the authors in [81] succeeded in theoretically predicting that these ranges can be partially overlap; i.e., they predict the possibility of designing the metamaterial of the above type with a negative refractive index in the visible spectral range. The figure of merit depends on the material parameters used in the calcula tions and reaches rather large values of 210. In view of its structure (polycrystalline matrix, spherical particles with a random arrangement or arrangement at sites of the cubic lattice), the metamaterial of this type is a bulk isotropic optical material. In [82], the authors considered a random mixture of spherical particles of two types, namely, SiC particles that provide a negative effective permeability of the mixture and gold particles that ensure a negative per mittivity. The results of the calculations performed for the particles with radii rSiC = 65 nm and rAu = 10 nm and volume fractions fSiC = 0.3 and fAu = 0.46 demonstrated that, in the wavelength range with a width of 20 40 nm near a wavelength of 370 nm, this bulk isotropic metamaterial should have a negative refractive index n = 14 and be characterized by high values of the figure of merit 2.5. It should be noted that the sys
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

533

tems consisting of spherical particles of two types (dielectric and metallic) were previously investigated theoretically [84] under conditions that particles of each type form a simple cubic lattice. The fulfillment of this condition in practice seems to be complicated. Therefore, at present, the problem associated with the design of bulk isotropic metamaterials with a neg ative refractive index has been on the agenda. An increase in the figure of merit of metamaterials is also an acute problem. In conclusion, we note that we did not concern the methods for fabricating the aforemen tioned metamaterials. These methods are described in the review [85]. Nowadays, the problems of materials science of metamaterials have gone far beyond the problem regarding the preparation of negative refractive index metamaterials. In particular, the problems considered in the next section require the fabrication of spatially inhomogeneous anisotropic metamaterials with con trolled electromagnetic properties. In a number of cases, these metamaterials have been produced using traditional techniques, such as the incorporation of metallic elements (for example, split ring resonators) into the unit cell of the metamaterial. In addition, non traditional approaches have also been proposed for designing metamaterials. As an example, Smolyaninov et al. [86] proposed to emulate anisotropic metamateri als by waveguides with a variable cross section. METAMATERIALS AND THE PROBLEM OF IMPLEMENTATION OF INVISIBLE OBJECTS A human imagination has long appealed to the idea of the possibility of creating invisible objects. The case in point was both objects that do not perturb the light propagation (for example, invisible man) and devices that cloak objects and make them invisible (cap of darkness). The implementation of invisibility or low observability of objects has been the subject of exten sive physical and engineering investigations for many decades. In order to reduce the observability of an object in scattered or reflected light, absorbing screens (see, for example, [87]) or antireflection films (see, for example, [1, 88]) have been used in a number of appli cations. The possibility of controllably changing the prop erties of metamaterials (see the previous section) resulted in the appearance of two approaches to the solution of the problem of invisibility of objects. The first approach [89] considers objects that have sizes smaller than or of the order of the wavelength and are built up from conventional materials and metamateri als so that the action of this object on the propagation of the electromagnetic field would be minimum (transparent objects). According to the second approach [90, 91], metamaterials should be used for
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

fabricating a cloak that would bypass electromagnetic radiation around an object and retain a constancy of the field outside the cloak. Ideas of the Design of Transparent Objects with Sizes Smaller than or of the Order of the Wavelength Al and Engheta [89] considered the possibility of applying plasmonic or metamaterial coatings in order to considerably reduce radiation scattering from spherical objects with diameters smaller than the radi ation wavelength. It should be noted that this work marked the beginning of an active discussion of the possibility of using metamaterials to solve the problem of the design of invisible objects. Before proceeding to the discussion of [89], we note that some ideas proposed in this work were previ ously put forward by other authors. First, let us dwell on earlier works concerning the problem of invisibility of objects without pretending to the completeness of a review. Initially, these works were associated with the investigations of anomalous light scattering in glasses. An analysis of investigations in the review [92] revealed an anomalous character of light scattering in phase separated glasses, i.e., glasses that undergo phase separation: in a number of cases, the depen dence of the scattering coefficient sc on the wave length has the form sc 8 (and differs from the Rayleigh dependence sc 4) and the scattering indicatrix is directed backward. The qualitative explanation of the anomaly in the angular dependence of the scattered radiation at the early stages of the diffusion limited phase transforma tion was proposed by Walker and Guinier [93]. It was assumed that the average refractive index of the system consisting of a particle and a diffusion zone around it is equal or close to the refractive index of the matrix, which decreases the scattering intensity in the limit of small scattering angles (small scattering vectors or large wavelengths). For larger angles (smaller wave lengths), the interference of fluxes scattered by differ ent volume elements leads to an increase in the scat tering intensity. In other words, in the case where the light wavelength considerably exceeds the particle size, the observability of the composite particle (parti cle and diffusion shell) in the matrix should be signifi cantly lower than that of a homogeneous particle. The quantitative theoretical analysis of light scat tering from a system consisting of a particle plus a shell in a matrix was performed in [94] (see also the review [92, formulas (12)(15)]). In [94], the author consid ered a homogeneous particle with the radius R1 and a homogeneous surrounding concentric spherical layer with the outer radius R2 in a homogeneous matrix. It was assumed that the permittivities of the particle 1, the spherical layer 2, and the matrix m obey the ine quality 1 > m > 2 (or the inequality 1 < m < 2). The
2010

534

ZHILIN, SHEPILOV

use of the RayleighGans approximation and the bal ance condition (compensation condition) [92, 94] R1 ( 1 2 ) = R2 ( m 2 )
3 3

(10)

for fairly long wavelengths ( > 4 m R2) leads to the spectral dependence of the scattering coefficient sc 8 and backward scattering indicatrix. This behavior is associated with the interference of waves scattered by the particle and the shell, which results in a decrease in the light scattering intensity in the range of long wavelengths due to the suppression of the dipole component (4) in the scattering. In this case, the scattering coefficient for the particle in the shell is sub stantially smaller than that for the particle directly embedded in the matrix. Therefore, the shell that sat isfies the compensation condition (10) considerably decreases the observability of the particle in the matrix. In the RayleighGans approximation, the compensation condition is the condition of weak observability and determines the values of the param eters for which the effective refractive index of the par ticle in the shell is equal to the refractive index of the matrix (we do not assign a rigorous physical meaning to the notion of the effective refractive index). Shatilov [94] used the simplest (rectangular) pro file of the dependence of the permittivity (refractive index) on the distance between the particle center and the point of observation. Goldstein [95] considered a more complex case when the refractive index in the shell is determined by the concentration profile of the component diffusing to the particle or from the particle. The conclusions obtained in [95] for the character of light scattering at early stages of phase separation are in agreement with those obtained in [94]. The RayleighGans approximation was used in [94, 95]. One of the conditions for its applicability, i.e., the closeness of the refractive indices of the matrix, particle, and shell [83, Chapter 6], frequently does not hold true. In this respect, the study of light scattering from particles in shells was continued; in this case, the term invisible was brought into use [96]. In the electrostatic approximation, Kerker [96] derived the relationship for the intensity of radiation scattered by a small nonabsorbing ellipsoid that con sists of an ellipsoidal core and a confocal shell and is embedded in a nonabsorbing matrix. By using this relationship, the author found the condition under which the scattered field disappears; i.e., the ellipsoid becomes invisible. In a particular case of a particle consisting of a spherical core in a spherical shell, the condition for the absence of the scattered field (the condition for the invisibility of the particle) has the form
3 R 1 = ( m 2 ) ( 1 + 2 2 ) R 2 ( m + 2 2 ) ( 1 2 )

(11)

(see formulas (28), (25), and (26) in [96] or Section 5.4 in the monograph [83]), where the used designa tions were introduced above. This relationship can be used in the cases where the field of the incident wave can be considered uniform within the particle; i.e., if the particle is fairly small as compared to the wave length. The quantitative condition of the smallness for the composite particle was not formulated in [83, 96], whereas this condition for the nonabsorbing homogeneous particle with the refractive index n and the radius R in the matrix with the refractive index nm is reduced to the fulfillment of two inequalities (see Section 5.2 in [83]): x 1 and mx 1, where x = 2R/m = 2Rnm/ is the diffraction parameter of the particle; m = n/nm is the relative refractive index; and and m are the light wavelengths in vacuum and matrix material, respectively. It should be noted that, when the permittivities of the particle, shell, and matrix are close to each other, formulas (10) and (11) lead to close ratios R1/R2. For a sufficiently small particle characterized by a signifi cant contrast of the permittivities of the particle, shell, and matrix, it is necessary to use condition (11), which in actual fact is the condition for the vanishing of the polarizability of the particle in the matrix (see Section 5.2 in [83]). It is also interesting to note that, by analyzing the possibility of satisfying the conditions for the invisibility of the particle in vacuum (formula (11) with m = 1), Kerker [96] considered a wide range of permittivities of the core and shell [, ], including the untraditional range of negative values and values smaller than unity. In [97], the authors analyzed the cross section of light scattering from a spherical particle that consists of a core and a shell and cannot obey the smallness cri terion. It was noted that, in the electrostatic approxi mation [96], this small composite particle is treated as a dipole so that the core and the shell are characterized by different polarizations. When condition (11) is ful filled, the polarizations of the core and the shell are equal in magnitude and opposite in sign; i.e., the dipole moment becomes equal to zero and the particle in the dipole approximation does not scatter light (is invisible). In [97], the discussion is based on the expansion of the scattering cross section in powers of the wave vector, which is valid for any finite particle size. The numerical calculations were carried out for the values of nm = 1.00, n1 = 1.97, and n2 = 0.66 (the refractive indices of the medium, core, and shell, respectively) and two diffraction parameters of the particle x2 = 2R2nm/ = 0.3 and 1.6, which character ize the particle size with respect to the wavelength. The scattering cross section was calculated as a function of the volume fraction of the core in the particle f = (R1/R2)3. For x2 = 0.3, the scattering cross section exhibits a sharp minimum at f1 0.415. The minimum scattering cross section for the composite particle is six orders of
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

535

magnitude smaller than that for the homogeneous particle that has the same size and consists of the material of the core or the shell. Therefore, the small composite particle for some ratio between the param eters becomes almost invisible. The obtained value of f1 for which the scattering cross section is minimum almost coincides with the value of f0 0.414 calculated in the electrostatic approximation (11) and depends only on the quantities nm, n1, and n2. In the case of x2 = 1.6, the scattering cross section is minimum at f2 0.31, which differs noticeably from f0. Moreover, the minimum in this case is not so deep; the cross section decreases by approximately two orders of magnitude as compared to the case of the homogeneous particle. Furthermore, in [97], it was also demonstrated that, when condition (11) is satisfied, the dipole com ponents in scattering are absent and the residual scat tering cross section is proportional to 8 for fairly long wavelengths. In this case, the predominant back scattering takes place. These inferences correlate with those previously obtained in the RayleighGans approximation [94]. The most significant qualitative conclusion follow ing from the results obtained in [97] is that the sub stantial decrease in the observability is possible only for small composite particles with sizes smaller than or of the order of the wavelength. In [98], the rigorous solution [99] to the problem of light scattering from a spherical particle consisting of a homogeneous core and a homogeneous shell was used to examine the cases in which the real part of the per mittivity of the shell or the core is negative. As a result of the rigorous solution [99] (see also [83, Chapter 8]), the scattering efficiency (the ratio between the scatter ing cross section and the cross sectional area of the particle) can be represented in the form of the series Q sc = ( 2/x 2 )
2

n=1

( 2n + 1 ) ( a

2 n

+ b n ),

(12)

in which the terms correspond to different multipole components of scattered radiation, beginning with the dipole component (n = 1), an and bn are the scattering coefficients, and x2 = 2R2nm/ is the size parameter introduced above. It was noted in [98] that, for small particles (at x2 < 0.3), all scattering coefficients, except for the coefficient a1 can be ignored in expansion (12). In turn, the coefficient a1 can be represented in the form of the expansion in powers of x2 with the main term 2 3 a 1 ix 2 3 ( ) ( + 2 2 ) + q 3 ( 2 2 + m ) ( 1 2 ) 2 m 1 , ( 2 + 2 m ) ( 1 + 2 2 ) + q 3 ( 2 2 2 m ) ( 1 2 )
GLASS PHYSICS AND CHEMISTRY Vol. 36

(13)

where we used the above introduced designations for the permittivities and the geometric factor q = R1/R2 (0 q 1), which represents the ratio between the radii of the core and the shell. In this case, the scattered field corresponds to the field of the oscillating dipole with the polarizability proportional to a1 (electrostatic approximation). The condition for anomalously low scattering, i.e., the vanishing of the numerator in expression (13) [98], is equivalent to the above derived condition (11). In the examples considered in [98], it was assumed that the particle is in vacuum (m = 1). In these exam ples, the authors calculated the efficiencies of scatter ing and absorption from the rigorous formulas (for mula (12) for the scattering efficiency) with exact expressions for the scattering coefficients. The most interesting is a series of the calculations with the use of the experimental data on 2 for a silver shell containing a dielectric core (1 > 1). Apart from the variation in the geometric factor q = R1/R2 in these calculations, the authors considered different wavelengths, different particle sizes, and different permittivities of the core. In particular, the scattering and absorption efficiencies were calculated as a function of the wavelength for particles with a specified geometry and composition. Although, in a number of cases, the scattering effi ciency for small particle can reach very low values (for example, at a specific wavelength for particles with a specified geometry), the absorption efficiency in these cases considerably (by several orders of magnitudes) exceeds the scattering efficiency. Consequently, under specific conditions, the scattering of monochromatic light from small particles that consist of a dielectric core and a silver shell in vacuum can be reduced con siderably, whereas the absorption of light by the same particles remains at a high level. The results obtained in [98] demonstrate that, in order to fabricate small particles invisible in vacuum, it is necessary to develop the shell material character ized by the negative real part and fairly small imaginary part of the permittivity. In [98], the authors also noted (but not considered) the possibility of using a material with the permittivity lower than unity. Therefore, the modern work by Al and Engheta [89] on the problem of invisibility of small particles in vacuum was preceded by a number of works, at least, a part of which we cited above. The investigation per formed in [89] is based on the analysis of light scatter ing from a particle that consists of a core and a shell in the cases where (meta)material of the core and/or the shell is characterized by negative permittivity and/or a negative permeability. This analysis was carried out in [100] (see also Erratum to this work [101]), where it was noted, in particular, [100, p. 2] that the solution obtained previously in [99] for a particle from conven tional materials is also applicable in the case under consideration when the signs of the expressions j j and j / j are chosen correctly. It should be
2010

No. 5

536

ZHILIN, SHEPILOV

noted that this choice in the case of j < 0 and j < 0 was first discussed by Veselago [3] and the light scattering in the case where only one of the permittivities of the particle (1 or 2) is negative was studied earlier in [98]. Al and Engheta [89] used a scattering series for a particle consisting of materials and/or metamaterials [99] that, with due regard for the remarks made in the previous paragraph, is equivalent to series (12). First and foremost, the authors in [89] derived the approxi mate relationship for the coefficients of the scattering series in the case of a small particle and demonstrated that, if the core of the particle is a dielectric, the main contribution to the scattering is made by the TM dipole mode (dipole mode of the electric type [83, TM Section 4.3.2]) with the scattering coefficient a1 or c 1 according to the designations of formula (12) and the work [89], respectively. By equating this coefficient to zero, the authors in [89] obtained the invisibility condi tion equivalent to relationship (11). Then, the authors performed the numerical calcu lations of the scattering series as a function of the geo metric factor q = R1/R2 for a number of values of the parameters R2/ (or R1/), j, and j (j = 1 for the core and 2 for the shell) under the condition that one of the components of the particle (the core or the shell) is a conventional dielectric (j > 1, j = 1) and the other component is an unconventional material or a metamaterial (j ' < 1, j ' 1). In the graphical repre sentation of the results of the calculations, apart from the total scattering cross sections, the values of indi vidual terms of the scattering series are also presented in [89], which allows one to judge the contributions of different normal modes [83, Section 4.3.2] to the scat tering. In [89], the results of the numerical calculations showed that the particle with a fairly small size (R2/ < 1/10) in which the core or the shell consists of the material with < 1 becomes unobservable when con dition (11) is fulfilled due to the suppression of the electric dipole component in the scattering. For parti cles with a larger size (R2 /5), noticeable scattering is determined by the magnetic dipole term (term in formula (12)) including the scattering coefficient b1 = c 1 ; this term and, hence, the observability of the par ticle can be significantly decreased by using the metamaterial with < 1 [89]. Therefore, the use of val ues < 1 provides additional possibilities of decreasing the cross section in the case where the particle is not sufficiently small. Moreover, the results of the calculations in [89] also demonstrated that the presence of losses in some limits does not strongly affect the invisibility. Finally, the authors assumed that the invisibility effect is not very sensitive to the shell shape, because the ability to sup press one multipole component in the scattering is not an exclusive property of a spherical shell.
TE

Zhou and Hu [102] used the Maxwell Garnett effective medium theory [103] and its improved vari ant [104] for the evaluation of the scattering properties of a composite particle that contains the material with < 0 and has a size considerably smaller than the radi ation wavelength. The estimates obtained agree well with the results of the calculations carried out using the rigorous scattering theory. In [102], the authors introduced the notion of the neutral inclusion, i.e., the composite particle with the effective permittivity equal to the permittivity of the environment, and showed that the spherical neutral inclusion consisting of the core and the shell satisfies the invisibility condi tions (11). The calculations performed in [102] dem onstrated that the neutrality condition rather accu rately describes the conditions under which the scat tering is minimum and, at the same time, is fairly simple even for particles with a complex shape and structure. The authors made the inference that the effective medium approximation and the notion of the neutral inclusion are useful for designing small transparent objects. In [105], the numerical simulation was used to investigate the light scattering from a core surrounded by a spherical shell from a plasmonic material. Specif ically, the calculations demonstrated that the transpar ency effect is stable with respect to rather substantial changes in the shape and electromagnetic properties of the core. The representation is accompanied by color illustrations of the structure of the calculated fields and movies. The light scattering from ensembles of two and four particles that consist of cores covered by shells from metamaterials or plasmonic materials was theoreti cally investigated in [106] by using the numerical sim ulation. The results of the calculations showed that the scattering cross section decreases considerably in the appropriate choice of the shell design, even though the size of the ensemble was noticeably larger than the electromagnetic radiation wavelength. Al and Engheta [107] proposed the idea of the fabrication of complex spherically symmetric particles consisting of a core and several shells as objects that are invisible in the use of electromagnetic radiation with several frequencies. As an example, the authors theoretically considered the case of the dielectric core that has a radius of 100 nm and is surrounded by two shells from different plasmonic metamaterials and showed that the use of the bilayer shell makes it possi ble to sharply reduce the scattering for the radiation with two wavelengths (500 and 625 nm in the example under consideration) irrespective of the angle of observation and the polarization of the incident beam. The scattering cross sections decrease by several tens of times for these two wavelengths as compared to the scattering cross section of the core; in this case, mod erate losses do not substantially affect the scattering level. Furthermore, two minima in the dependence of
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

537

the scattering efficiency on the wavelength are not sharp, which suggests that the phenomenon does not exhibit a pronounced resonant character and the core in the shells almost does not scatter light in two wave length ranges. On the whole, the authors made the inference that the use of multilayer plasmonic shells allows one to drastically reduce the light scattering from a dielectric particle at several wavelengths in the optical range. It was noted that the results can be directly extended to the range of lower frequencies by fabricating plasmonic shells from metamaterials, such as those considered in [108]. Since the perturbation of the incident field for several wavelengths of invisibility takes place only in the near field zone of the particle, the invisibility conditions should be fulfilled for an ensemble of identical particles for which the total size can be larger than the wavelength [106]. It should also be noted that, to the best of our knowledge, data on attempts to experimentally imple ment objects that are invisible in vacuum and have sizes smaller than or of the order of the electromag netic radiation wavelength are not available in the lit erature. It seems likely that, from the experimental viewpoint, we can only say on small sized objects in matrices. Examples can be provided by particles sur rounded by diffusion zones in the course of phase sep aration in glasses. For the particle plus zone system, the electric dipole component in the light scattering is absent, which leads to anomalously weak light scatter ing in these systems [92]. Theoretical Development and Attempts to Implement Invisible Cloaks from Metamaterials that Conceal the Objects Contained in them from an External Observer Progress in the development of metamaterials have become a prerequisite for the statement of the problem of the design of invisible cloaks that ensure the invisi bility of the objects contained in them from an external observer. Two works [90, 91] in which methods for transforming coordinates (mappings methods) were used for constructing theoretical models of these cloaks were simultaneously published in June 2006. Leonhardt [90] primarily noted that, according to the results obtained in [109, 110], a complete invisibil ity cannot be achieved because of the wave nature of light. This inference follows from the theorem of the uniqueness of the solution to the inverse scattering problem [109], according to which the scattering data completely determine the spatial refractive index pro file. In this respect, data of light scattering from empty space can correspond only to empty space and an ideal invisibility is impossible. However, as was noted in [90], this theorem does not impose quantitative limi tations on the degree of nonideality and, moreover, cannot be applied directly to light rays, i.e., to the light propagation under the conditions of geometrical optics. The work [90] is devoted to the problem of the
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

prediction of the properties of a medium that provides an ideal invisibility in the geometrical optics limit (see also [111] in which some aspects of the problem are considered in more detail). In [90], the author considered the case where the refractive index n is uniform along some direction, i.e., depends only on the (x, y) coordinates in the plane perpendicular to this direction, and the wave vector of the incident plane electromagnetic wave lies in this plane. Then, the problem is reduced to a two dimen sional problem. Moreover, in [90], it was assumed that the refractive index does not change substantially at distances of the order of the wavelength and the geo metrical optics approximation was used. By introduc ing complex numbers z = x + iy for the description of the coordinates in the (x, y) plane and using conformal mapping, the author found the refractive index profile that provides the restoration of the initial direction of propagation of light rays at the exit from the region of refractive index nonuniformity (Fig. 13); in this case, the rays do not enter into some spatial region. The examples of the restoration of the initial direction of propagation of light rays for fluxes of two directions are shown in Figs. 13a and 13b. The dark spot at the center of the figure indicates the region into which light rays do not enter. The propagation of incident light rays does not depend on the properties of this region, and the object with appropriate sizes placed in it, as well as the refractive index nonuniformity itself, will be invis ible for the external observer, because the rays restore the propagation direction after the passage of the non uniformity. Therefore, the found profile represents an invisibility device. It is important to note four specific features of the invisibility device considered in [90]. First, the refrac tive index of the device is isotropic (in the example considered in [90], its value varies from 0 to 36). Sec ond, since the found refractive index profile does not possess a cylindrical symmetry, the character of prop agation of the flux of light rays depends substantially on the direction of their propagation (Fig. 13). Third, there is a jump in the refractive index on the cylindrical surface with the circular cross section |z| = a, which illustrates the propagation of the central ray in Fig. 13a (where a is a fixed real parameter of the model). Fourth, in the strict sense, the size of this device is not limited because the refractive index profile at |z| > a is determined by the expression n(z) = |1 (a/z)2|, i.e., differs from unity for any finite value of z. Leonhardt [90] called attention that the imperfec tion of the device is associated with the partial reflec tion of light from the surface |z| = a and with the time delay of rays that bend around the object. He assumed that the reflection can be reduced by introducing a transition layer instead of a sharp boundary or using antireflection films. In his opinion, the imperfections of the invisibility device in rays can be made as small as
2010

538

ZHILIN, SHEPILOV

Fig. 13. Trajectories of rays in the Leonhardt invisibility device [90] in the cases of propagation of the incident flux (a) in the direc tion of the x axis and (b) at an angle with respect to this axis. The black circle at the center of the figure indicates the invisibility region.

the inaccuracy of geometrical optics, whereas the object cannot be completely hidden from waves. Some aspects of the invisibility device of the above type, including the problem of the time delay, are con sidered in more detail in [111]. The method proposed in [90] did not receive wide use. However, the ideas formulated in [90] were used to develop the model [112] in which the authors eliminated the imperfec tions of the device (time delay and reflection effects). Unlike Leonhardt, who discussed the invisibility problem in the geometrical optics approximation, Pendry et al. [91] used the Maxwells equations in their approach to the theoretical treatment of the invisibil ity cloak. These authors were based on the results obtained in [113] (they are reproduced in the elec tronic appendix to [91]), according to which the Max wells equations for monochromatic radiation in the absence of free charges and currents are invariant with respect to spatial deformations, i.e., transformation of coordinates, if this is accompanied by the corre sponding renormalization of the permittivity and permeability . It should be noted that, in the general case, this renormalization leads to coordinate depen dent tensors of these quantities, even though the per mittivity and permeability in the initial coordinates were homogeneous and isotropic. If any region of the space is not deformed, i.e., the initial and transformed coordinates of points in this region coincide with each other, the permittivities, permeabilities, and electro magnetic fields of the initial and transformed systems also coincide within this region. Generally speaking, the electromagnetic fields in the initial and trans formed systems differ and are related by particular expressions.

In order to more clearly illustrate these statements, we demonstrate them using a simple transformation as an example. Let us write the Maxwells equations H = 0 E/t, E = 0 H/t (14) in free space ( = = 1, the initial system) in the spherical coordinate system (r, , ) [114]. For exam ple, for the component of the first equation in system (14), we obtain (this component is chosen as the most compact in the writing) E 1 [ rH ] H r . = 0 r r t (15)

Now, we perform the transformation of the coordi nates R = R(r) that does not include the angles, where R(r) is the one to one function and R(r)
r

r.

(16)

By changing over to the variable R in Eq. (15) with allowance made for the differentiation rule ()/r = (dR/dr)()/R and performing simple manipula tions, we obtain the following equation for the trans formed system: 1 R r H 1 H r R R R dR/dr = 0 1 r E . dR/dr t R

(17)

Equations (15) for the initial system and (17) for the transformed system are identical in the form but differ in the values of the component of the permittivity:
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

539

= 1 for the initial system and ' = (dR/dr)1 for the transformed system. Equation (17) also indicates that the field components in the transformed system are represented by the expressions enclosed in parenthe ses, i.e., expressed through the corresponding compo nents in the initial system. The transformation of other components of the first and second equations of system (14) can be con sidered in a similar manner. The equations for the transformed system are the Maxwells equations in the medium with the permittivities and permeabilities 'R = 'R = r dR , 2 R dr
2

permittivities and permeabilities in the transformed system (20) in the region R1 < R < R2: 'R = 'R = R 2 R R 1 2 , R2 R1 R

(21)

R2 ' = ' = ' = ' = . R2 R1 Therefore, if the cloak, i.e., the spherical layer R1 < R < R2 with the material that satisfies expressions (21) at a specified frequency of radiation, is placed in empty space, this cloak does not perturb fields in the spatial region R > R2 (i.e., external fields) with the specified frequency irrespective of the objects located inside the cloak (in the region R < R1). These fields have the form identical to that in the initial system in which the cloak and its content are absent. The fabri cation of this invisible cloak would mean the possibil ity of hiding the objects contained inside the cloak for the external observer. In this case, the space is sepa rated into two regions R > R1 and R < R1, which are not electromagnetically related to each other for the radi ation with the frequency under consideration. Pendry et al. [91] noted that the cloak material at the outer boundary (r = R2) obeys the relationships ' ' ' ' ' ' = = 1/ R and = = 1/ R , which ensure the absence of reflection at the boundary. It should also be noted that the anisotropy and spatial inhomo geneity of the permittivities and permeabilities are fun damentally important properties of the cloak material and can be provided for metamaterials [115117]. According to [91], the possibility of obtaining small values of the quantities ' and ' seems to be more problematic. In order to demonstrate the operation of this cloak, the authors considered two examples. In the first example, it was assumed that the radiation wave length is small R2 and the path of rays that are inci dent on the cloak from an infinitely distant source and characterize the behavior of the Poynting vector is cal culated in the geometrical optics approximation. Fig ure 14 presents the results of these calculations in the plane that is parallel to the direction of incident flux and passes through the center of the sphere (Fig. 14a) and in the three dimensional representation (Fig. 14b). It can be seen from this figure that the rays enter into the cloak and bend around the cloaked region R < R1, so that the initial direction of ray prop agation is restored at the exit of the cloak; i.e., the con tent of the cloaked region and the cloak itself are invis ible for the external observer. In the second example, R2 (near field the authors considered the case limit), placed a point charge in the vicinity of the cloak, and used the electrostatic approximation for constructing the field lines of the electric induction vector (Fig. 15). It can be seen from Fig. 15 that the field does not penetrate into the cloaked region and
2010

(18)

1 ' = ' = ' = ' = . dR/dr In this case, the field components in the transformed system ( H , E ) and the initial system (H, E) are related by the expressions HR = 1 H r, dR/dr r H = H , R r H = H R (19)

(the electric field components are transformed in a similar way). Formulas (18) and (19) suggest that, as was noted above, the permittivities, permeabilities, and electromagnetic fields in the initial and trans formed systems coincide in the spatial regions that do not undergo deformation (R = r). In [91], the authors considered the transformation R = R 1 + r ( R 2 R 1 )/R 2, ( r < R 2 ); R = r, ( r < R 2 ); ' = ; ' = , (20)

which corresponds to compression of a sphere of radius R2 into a concentric spherical layer with the outer radius R2 and the inner radius R1. This transforma tion represents a particular case of the transformation R = R(r) considered above (see relationships (14)(19)). Upon transformation (20), the spatial region r > R2 is not deformed (R = r); i.e., as follows from the formu las given in [113] or formulas (18) and (19), the per mittivities, permeabilities, and electromagnetic fields are identical for the initial and transformed system at R = r > R2. For example, if the plane wave propagates in the initial system, the field in the transformed system outside the sphere R > R2 will be also described by the plane wave. Then, the region R < R1 in the transformed system has no analog in the initial system, the permit tivities and permeabilities in this region can be speci fied arbitrarily, and the external field, i.e., the field of the initial system does not penetrate into this region. Finally, by using the formulas given in [113], as was done in [91], or the simpler relationships (18) and (19), we can derive the following expressions for the
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

540

ZHILIN, SHEPILOV

Fig. 14. Trajectories of rays in the invisibility cloak proposed in [91] and characterized by the permittivities and permeabilities defined by expressions (21) for R2 = 2R1. Trajectories are shown (a) in the plane that is parallel to the direction of the incident flux and passes through the center of the cloak and (b) in the three dimensional space.

that the field outside the cloak is identical to that in its absence. Finally, the authors asked the question as to whether the cloaking effect occurs for some frequency range or only for one frequency. In the example under consideration, the case in point was one fixed fre quency. It should be noted that, in terms of geometri cal optics (Fig. 14), the rays entering into the cloak propagate along curved, i.e. longer, trajectories as compared to those propagating in free space. However, at the exit of the cloak, their phase should be identical to that for rays propagating in free space, because the cloak does not perturb the external fields. This implies

Fig. 15. Point charge in the vicinity of the invisibility cloak [91]. Field lines of the electric induction vector are shown under the assumption that R2 (near field limit). The field does not penetrate into the cloaked region and, out side the cloak, is identical to that in the absence of the cloak.

that the phase velocity of rays in the cloak is higher than that in free space and this statement violates no physical laws. By assuming that the dispersion is absent, the phase and group velocities should be iden tical, whereas the group velocity cannot exceed the velocity of light. Therefore, the dispersion should take place and the cloak can efficiently operate only at one frequency. The theorem of the uniqueness of the solution to the inverse scattering problem [109] was not men tioned in [91], and, naturally, the fact that the results obtained contradict this theorem was not commented. The corresponding comment was done in more recent work [118] and reduced to the fact that the uniqueness theorem [109] was proved for an isotropic media and inapplicable to the case of the cloak from a substan tially anisotropic material [91]. According to Rozanov [119], the anisotropy of the cloak material is of little significance for the violation of the uniqueness theorem. This author proposed another explanation for the violation of this theorem, which is based on the fact that the permittivity and the permeability tend to infinity at the inner bound ary of the cylindrical cloaking cloak (see below formu las (23)). This point of view seems to be incorrect. Actually, in the case of the spherical cloak, the permit tivities and permeabilities (relationships (21)) remain finite and the only existing explanation is associated with the anisotropy of the cloak material. Furthermore, it is improbable that the factors responsible for the violation of the uniqueness theorem in the three dimensional and two dimensional cases are different. Finally, reasoning from formulas (18) and expressions (22) derived below, we can predict invisible optical inhomogeneities with smoothly varying finite (but anisotropic) permittivities and permeabilities in both the three dimensional and two dimensional cases. For these inhomogeneities,
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

541

which are not invisibility cloaks, the uniqueness theo rem is also violated. In this respect, the anisotropy of the material should be considered the fundamental factor responsible for the violation of the uniqueness theorem. Let us briefly compare the invisibility cloak models proposed in [90, 91]. The approach proposed by Leonhardt [90] can be directly used only in the two dimensional case, whereas the approach by Pendry et al. [91] developed for the three dimensional case can be easily extended to the two dimensional case (see below). The cloaks considered in [90] can strictly ensure invisibility only in the geometrical optics limit and require additional measures for the elimination of reflection effects at the boundaries of the outer and inner regions, whereas the cloaks proposed in [91] are free of these limitation and complication. For the cloaks proposed by Leonhardt [90], the size of the region in which the refractive index differs noticeably from unity, i.e., the outer size of the cloak, is consider ably larger than the size of the invisible region, whereas, for the cloak proposed by Pendry et al. [91], there is the only limitation, namely, the cloak radius should be larger than the radius of the invisibility region. The cloaks proposed in [90] do not possess a circular symmetry and consist of an isotropic material, and the cloaks proposed in [91] are spherically sym metric structures from an anisotropic material. On the whole, the fabrication of the cloaks proposed in [91] seems to be more promising, which is reflected in the subsequent development of the invisibility idea in the literature. The approach by Leonhardt [90] is of inter est in that it shows the possibility of creating the invis ibility device with the use of the isotropic material, at least, for the two dimensional case in the geometrical optics approximation. It is interesting that, already in 1961, Dolin pub lished the paper [120] in which he very closely came to the modern solution to the invisibility problem. First, Dolin investigated the invariance of the Maxwells equations with respect to a specific type of transforma tions of the space metric and permittivities and perme abilities of a medium, which was also done in [113] but considerably later, and proposed to use this invariance for the solution to some electrodynamic problems. Second, as one of the applications, Dolin considered the transformation of free space upon changing over from the spherical coordinates r, , and to the coor dinates R(r), , and , where R = R(r) satisfies condi tion (16), and found the type of the inhomogeneity that does not perturb the external electromagnetic field, i.e., is invisible for the external observer. In order to predict the invisibility cloak in its modern meaning, he was short of one step, namely, the use of
1 It

the transformation R(r) that transforms the complete space into its part, such as transformation (20). The ideas formulated in [91] were subsequently used in [121, 40] for the theoretical development of simpler (simplified) cloaks. In this case, for simpli fication, the authors examined the two dimensional case (cylindrical cloaks), radiation with a particular polarization, and, as was noted more recently [122, 123], rejected some terms in the wave equation. In the analysis of a cylindrical cloak, it is conve nient to use the cylindrical coordinate system (r, , z), where r is the distance from the point of observation to the z axis (in contrast to the aforementioned case of the spherical cloak in which the same designation was used for the distance from the point of observation to the origin of the coordinates). We choose free space with the coordinates (r, , z) as the initial system and consider the transformation at which the variables and z remain unchanged and the variable r transforms into R = R(r) in a one to one manner, where R(r) sat isfies condition (16). As was done for the spherical cloak (expressions (14)(19)), we can demonstrate that the Maxwells equations (14) remain invariant with respect to the transformation under consider ation if the permittivities and permeabilities in the transformed system have the following form: r dR , 'R = 'R = R dr ' = ' = R 1 , r dR/dr

(22)

'z = 'z = r 1 . R dR/dr In the case of the linear transformation R(r) defined by relationships (20), the cylindrical cloak has the outer radius R2 and the inner radius (radius of the invisibility region) R1 and formulas (22) are reduced to the following formulas for the permittivities and per meabilities of the cloak material: 'R = 'R = R R1 , R ' = ' = R , R R1

R2 2 R R1 , 'z = 'z = R 2 R 1 R

(23)

( R 1 < R < R 2 ).

should also be noted that the permittivities and permeabilities obtained in [120] are inverse with respect to those following from formulas (18). GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

Cummer et al. [121] considered the incidence of a monochromatic polarized plane wave with the electric vector parallel to the z axis (TE polarization) on an ideal cylindrical cloak (formulas (23)) and a num ber of its modifications. For the polarization under consideration, the problem is determined by only three material functions, i.e., the permittivity 'z and the permeabilities 'R and ' . Moreover, the authors demonstrated that the propagation of radiation with the TE polarization in the cloak is determined by two material functions, namely, the products of the perme abilities and the permittivity 'R 'z and ' 'z . These
2010

542

ZHILIN, SHEPILOV

products for the ideal cloak (formulas (23)) are identi cal to those for simpler profiles proposed in [121]: '' = 1, R2 2 '' , z = R 2 R 1 R R 1 '' R = . (24) R
2

In this case, the authors noted that, although the phase front and the direction of the energy flux inside the simplified cloak described by expressions (24) are the same as in the ideal cloak (formulas (23)), the simpli fied cloak loses an ideality and there appears a reflec tion from the outer boundary. Cummer et al. [121] performed numerical calcula tions of the fields; in this case, the results of the calcu lations are approximate in character. The analysis of the results led the authors to the following conclu sions. (1) The cloaking effects are most pronounced for the ideal cloak (formulas (23)). In this case, the insig nificant scattered field is associated with the inaccu racy in the results of the calculations. (2) The losses in the cloak material reduce the cloaking effect; however, even for a loss tangent of 0.1, the cloaking remains rather effective. For a loss tan gent of 0.01, the influence of the losses becomes almost negligible. (3) In the case of a discrete cloak consisting of eight layers, which corresponds to the approximation of a ' continuous profile of the permittivity z and the per meabilities 'R and ' for the ideal cloak (formulas (23)) by a stepwise profile, there occurs an effective cloaking with a low reflection. Systematic perturba tions of the order of 10% for the permittivities and per meabilities do not disturb physics of cloaking. The authors assumed that the corresponding inference also holds true for random perturbations. (4) For simplified profiles of the permittivities and permeabilities (expressions (24)), there is a noticeable scattering in all directions. However, the distribution of the phase of the electric field in this cloak is in agreement with that obtained for the ideal cloak. The authors believed that the fabrication of the simplified cloak can be the first step toward the experimental demonstration of the cloaking phenomenon. The results obtained in [121] allowed one to make an effort to experimentally implement the invisibility cloak. Schurig et al. [118] fabricated and investigated the simplified invisibility cloak (expressions (24)) for radiation with a frequency of 8.5 GHz. In [118], the authors noted that, in the geometrical optics approxi mation, the paths of rays, as well as the angles of refraction at the boundaries, in the simplified and ideal cloaks are identical and that the presence of reflection at the outer boundary is the only disadvan tage of the simplified cloak. In order to fabricate the simplified cloak (expres sions (24)), the authors of [118] proposed to use an

inhomogeneous metamaterial in which the unit cell represents a copper split ring resonator on a dielectric substrate. A variation in the geometric parameters of the resonator makes it possible to vary the electromag netic response of the unit cell. The resonator axes were oriented along the radial directions. The unit cell size (3 mm) was considerably smaller than the operating radiation wavelength (3.5 cm), which allows one to consider the effective permittivities and permeabilities of metamaterials. In the radial direction, the cloak con sisted of ten layers. The inner and outer radii of the cloak were chosen to be R1 = 27.1 mm and R2 = 58.9 mm. The authors obtained the required values of '' and '' ( R ) z R (expressions (24]) as a result of the variation in two geometric parameters of the resonator as a function of the number of the layer that contains the resonator. The choice of the necessary values of these parameters was based on the numerical calculations of the charac teristics of scattering for the unit cell with the specified geometry and on the subsequent use of these charac teristics in the conventional procedure for determin ing the effective permittivities and permeabilities of the material [66, 124]. In [118], the authors noted that the optimization of the cloak by simulating the prop erties of its real structure determined by details of sev eral thousands of resonators seems to be impossible. The authors fabricated the designed cloak, experi mentally investigated its cloaking properties, and compared them with the calculated properties of the ideal cloak (formulas (23)) and simplified cloak (expressions (24)) for which the permeability '' ( R ) R was approximated by a piecewise constant function in order to reproduce the structure of the real cloak con sisting of ten layers. The measurements were carried out for a discrete set of frequencies, which involves the hypothetical operating frequency. The analysis of the results demonstrated that the optimum frequency is equal to 8.5 GHz, i.e., the frequency that was used in the calculations associated with the optimization of the unit cells of the metamaterial. The prepared cloak decreases both the backscattering (reflection) and the shadow effect. A comparison with the results of the calculations allowed the authors to make the inference that the results of the simulation are in agreement with the experimental data. However, the authors noted the presence of reflection, which is a consequence of the use of the simplified permittivity and permeability profiles. Reasoning from agreement between the theory and experiment, the authors in [118] drew the conclusion regarding the possibility of designing metamaterials that satisfy detailed and stringent requirements (including specified gradients of properties in a non rectangular geometry of the problem). They also noted that, although the achieved invisibility is not perfect due to the simplifications and absorption of the metamaterial, the experimental results obtained illus trate the mechanism of electromagnetic cloaking and
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

543

demonstrates the possibility of fabricating media described in the method of the transformation of coor dinates with the use of the metamaterial technology. In order to fabricate the simplified invisibility cloak (expressions (24)) proposed in [121], it is necessary to control the magnetic response of metamaterials, which, as was shown in [118], is possible in the giga hertz frequency range. Cai et al. [40] noted that the generation of the magnetic response in the optical spectral range is associated with the high losses [41, 42] and proposed a nonmagnetic variant of the simplified cloak, which, from the viewpoint of the authors, can be implemented for visible light. The ideal cylindrical cloak (formulas (23)) was chosen as the standpoint; however, the study was performed not for the TE mode, as in [121], but for the TM polariza tion, when the magnetic field is polarized along the z axis and the problem is determined by three material ' functions, namely, the permeability z and the per mittivities 'R and ' . In [40], the authors noted that the propagation of radiation with the TM polarization in the cloak is determined by two functions, i.e., the products 'R 'z and ' 'z , and proposed the simplified profiles '' = 1, z R2 , '' = R 2 R 1
2

cloak: f(R) = (R1/R)f(R1). This dependence is charac teristic of the case where particles uniformly fill seg ments R [R1, R2] in the radial directions. The authors performed calculations and demonstrated that the metamaterial of this type that consists of silver parti cles in the silicate matrix can be used for fabricating the cloak with the operating radiation wavelength of a heliumneon laser ( = 632.8 nm). Therefore, Cummer et al. [121] made the infer ence, which was supported in [118, 40], that the reflection at the outer boundary is the only disadvan tage of the simplified cloak. As was noted in [40] (see above), the reflection is associated with the mismatch between the impedances of the cloak and free space. In order to eliminate the undesirable reflection of the simplified cloak, instead of the linear transformation of the radius (20) in the cylindrical coordinate system, which was considered in [121, 118, 40], Cai et al. [125] proposed to use a transformation R(r) in a more gen eral form (higher order, for example, quadratic) that compresses the cylindrical region R < R2 into a cylin drical layer R1 < R < R2. An additional flexibility of this more general transformation makes it possible to match the impedance of the simplified cloak to the impedance of surrounding empty space. As a specific example, Cai et al. [125] considered the quadratic transformation with the fitting parameter and showed that the choice of a particular value of this parameter allows one to describe the permittivity and permeabil ity profiles of the nonmagnetic simplified cloak, which is characterized by the impedance Z = 1 at the outer '' '' boundary. For this cloak, three parameters z , , and '' are equal to unity at the outer boundary and are R matched to the parameters of surrounding vacuum. In this case, the radiation with the TM polarization not only bends around the cloaked region but also it is not reflected from the outer boundary of the cloak. Moreover, Cai et al. [125] carried out the numerical calculations of the fields upon incidence of radiation with the TM polarization on the ideal, simplified, and optimum simplified cloaks containing ideal metal cyl inders. By comparing the results of the calculations, the authors make the inference that the simplified lin ear cloak noticeably scatters light, whereas the scatter ing in the case of the optimum simplified cloak is neg ligible and not directional, which is an important property of the quasi ideal cloak. It is the authors opinion that the, in the geometrical optics limit, the optimum cloak is as perfect as the ideal cloak. Therefore, the principle of optimization of simpli fied cloaks with the use of general transformations R(r) and the optimized quadratic cloak, which were pro posed in [125], are one more step forward in the prac tical implementation of invisibility. Cloaks based on high order transformations were rigorously analyzed by Weder [126].
2010

R 2 2 R R 1 2 '' = , R R 2 R 1 R

(25)

for which these products are identical to those for the ideal cloak (formulas (23)). In this case, the only unde sirable effect in the use of the simplified profile is non zero reflection due to the mismatch of the impedance at the outer boundary of the cloak. The ideal cloak (formu las (23)) is characterized by the impedance Z = 'z / ' = 1 at the outer boundary (R = R2) and satisfies the condition of the perfect matching with vacuum, whereas the impedance for the simplified cloak (expres sions (25)) is Z = 1 R12, where R12 = R1/R2, and the reflection level can be estimated to be |(1 Z)/(1 + Z)|2 = [R12/(2 R12)]2. In [40], it was noted that the nonmagnetic charac ter of the simplified cloak (expressions (25)) allows one to eliminate the most complex problems associ ated with its fabrication. As a possible variant of the implementation of this cloak, the authors proposed to use the cylindrical dielectric layer with distributed identical pieces of the metal wire of subwavelength sizes (elongated particles) oriented along the radial directions. The ratio between the length of particles and their diameter is assumed to be specified, and their distribution in the matrix can be random to some extent and is characterized by the volume fraction f, which depends only on the distance to the axis of the
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

544

ZHILIN, SHEPILOV

A higher flexibility of high order transformations allows one to adapt the properties of particular struc tures for the use them as simplified cloaks. Cai et al. [127] proposed the concepts of two structures, calcu lated their effective properties, and demonstrated that these structures can be employed as simplified cloaks corresponding to quadratic transformations. The first system, i.e., a metaldielectric structure, is intended as a simplified nonmagnetic cloak for visible light with the TM polarization. The second system consisting of wires prepared from the polaritonic material SiC with air interlayers is promising as a simplified cloak for IR radiation with the TE polarization ( 13.5 m). Unfortunately, none of the cloaks is optimum; i.e., they should be characterized by noticeable scattering. The justification of the simplified cloak concept [121] came under criticism in [122, 123]. In [122, 123], the authors were based on the wave equation written for the zth component of the electric field inside the cloak and drew the conclusion that these equations for the ideal cloak (formulas (23)) and the simplified cloak (expressions (24)) could be equivalent only in the case if the permeability would be constant for each cloak. As can be seen from formulas (23), this quantity for the ideal cloak is variable. As a result, the authors made the inference that the wave behavior in the simplified cloak differs from that in the ideal cloak. A similar inference can be made for simplified nonmagnetic cloaks of linear (expressions (25)) and general types, because the sim plification concept is based on the same principle. This inference explains the presence of scattering from the optimum quadratic cloak [125]. The imperfection of the simplified cloak (expressions (24)) is illustrated by calculating fields in [122]. Yan et al. [123] proposed a variant of the simplified linear cloak that, like the optimum quadratic cloak, obeys the condition for the ideal matching to vacuum. The performed calculations of the fields showed that the characteristics of invisibility of this cloak are sig nificantly better than those of the cloak described by expressions (24) and are at approximately the same level as for the optimum quadratic cloak. However, as was noted by the authors in [123], if it is desirable to obtain a thin cloak, their cloak is more preferred than the optimum quadratic cloak, because the geometry of the latter system is limited by the condition R2 > 2R1. It should also be noted that Collins and McGuirk [128] proposed the technique for deriving the permit tivities and permeabilities of simplified cloaks, which makes it possible to substantially improve the charac teristics of invisibility. The fabrication of the invisibility devices requires the use of metamaterials with properties that are very difficult to achieve, if at all possible. Leonhardt and Tyc [129] proposed the approach that makes it possible to weaken these requirements and, according to the authors, can favor the solution to the problem of invis ibility over a wide spectral range.

Furthermore, attempts were made to theoretically justify the possibility of fabricating invisibility cloaks (strictly speaking, with a low observability) by using isotropic materials. Huang et al. [130] theoretically developed a layered cloak from isotropic materials, in which the layered material can be considered an effec tive medium with the anisotropy dependent on the radius. It was demonstrated that, for an appropriate choice of the permittivity, permeability, and layer thicknesses, the reflection at the boundary is low and the radiation bend around the cloaked region. A simi lar investigation was carried out in [131]. Sun et al. [132] used the geometrical optics approximation and found the refractive index profile that ensures a low scattering level and smooth bending of radiation around the cloaked object. The simulation of the fields revealed that the proposed structure is effective. According to the authors of [130, 132], their approaches are especially promising from the point of view of the fabrication of cloaks for the optical spectral range. The concept of invisibility devices proposed in [133] is somewhat beyond the title of this section. This concept, like those considered above, is based on the transformation of coordinates; however, the hidden object is located not inside the device but outside it. The two dimensional device represents a dielectric circular cylinder surrounded by a cylindrical layer from a material with a negative refractive index with antiobject located inside. The properties of the anti object (shape, location, permittivity, permeability) are determined by the properties of the object and speci fied so that the system (object + device) should be invisible for the external observer. The conclusions drawn regarding the possibility of occurring this type of invisibility by using the method for transforming the coordinates were confirmed by the results of numeri cal calculations of the fields. It should be emphasized once again that the antiobject and, hence, the device, as a whole, are unique and intended for a particular object. Apart from the work by Schurig et al. [118] in which the authors attempted to experimentally fabri cate the invisibility cloak, there are several works con cerning this problem. It should be immediately noted that all the devices described below do not ensure an ideal invisibility due to particular factors (simplified type of cloaks, incomplete correspondence of electro magnetic characteristics of fabricated metamaterials to required parameters, including the presence of losses). Kante et al. [134] used the metamaterial consisting of split ring resonators for the design of the invisibility cloak operating at frequencies of approximately 11 GHz. The invisibility cloak for the frequency range 2442 MHz was produced using an inductorcapaci tor network [135]. For a cylindrical cloak of a quite different type, which was fabricated from metal layers and had an outer diameter 2R2 = 7 cm and an inner
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

545 (b)

diameter (diameter of the cloaked region) 2R1 = 3 cm, the cloaking effects in the frequency range of approxi mately 3 GHz were demonstrated experimentally and using the results of the calculations [136]. The calcu lations performed by the authors for a similar nanoshell with diameters 2R2 = 226 nm and 2R1 = 100 nm revealed cloaking properties of this cloak for visible light with a wavelength of approximately 600 nm. As regards the visible light range, we note the results obtained in [137, 86]. Smolyaninov et al. [137] reported on the fabrication of the two dimensional structure with a decreased observability on the basis of metamaterials. The distribution of the radial compo nent of the permittivity in this structure is close to the distribution that is characteristic of a nonmagnetic cloak operating in the visible light range. As was noted above, the authors in [86] proposed to emulate aniso tropic metamaterials by waveguides with a variable cross section. By using this approach, the authors demonstrated the cloaking effects in the visible light range with the size of a cloaked region that exceeds the wavelength by a factor of 100. Li and Pendry [138] proposed and theoretically investigated a new strategy for cloaking of objects, i.e., hiding under the carpet, which was theoretically implemented in [139142]. This strategy consists in hiding a convexity on the reflecting surface from an external observer with the use of the corresponding device by emulating the reflection from a flat surface. The object (with appropriate sizes) invisible for the observer can be placed under this convexity. For sim plicity, the authors in [138] considered the two dimensional case when the fields and properties of the medium do not depend on the coordinate z and it is assumed that the electric field is polarized along the z axis and the reflecting material is an ideal conductor. The cross section of the invisibility device, the convex reflecting surface, and the object hidden under it (physical system) by the plane z = const is shown in Fig. 16a. The flat reflecting surface on which the homogeneous dielectric with the rectangular cross section is placed is chosen as a virtual system (Fig. 16b). The sizes of this rectangle are identical to the outer sizes of the invisibility device. In [138], the authors consider the transformation of coordinates that transforms the virtual system into the physical system and determine the permittivity and permeabil ity of the spatially inhomogeneous material of the invisibility device. If this device is implemented, its action on the external electromagnetic field will be the same as in the case of the homogeneous dielectric on the flat surface. Therefore, the imperfection of the sur face will be cloaked for the external observer. As was shown in [138], this device can be fabricated with the use of materials with a nonsingular permittivity and permeability and, what is more, almost isotropic materials, which is an important advantage. The first attempt to experimentally implement this device was
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

(a) 1 Physical system y x 3 2 5 h Virtual system 4 0

Fig. 16. Cross sections of (a) the physical system consisting of (1) the invisibility device, (2) the convex reflecting sur face, and (3) the object hidden under it and (b) the virtual system representing (4) the flat reflecting surface and (5) the rectangular homogeneous dielectric strip atop of the surface (according to [138]).

made by Liu et al. [139]. In order to provide the nec essary complex spatial distribution of the permittivities and permeabilities in the device, the authors designed the metamaterial consisting of several thousands of structural elements (unit cells). The unit cell 2 mm in size contained an I shaped metal element with sizes that were varied as a function of the position of the unit cell and determined using the automated design pro cess. These elements are nonresonant, which ensures the cloaking effect of the device over a wide operating range of frequencies (from 13 to 16 GHz in the per formed experiment) and extremely low losses. The authors believed that the scaling of this device (an increase in the operating frequency due to the decrease in the size of the unit cells of the metamate rial) allows one to go toward optical frequencies. However, at present, the problem of the fabrication of devices of the type hiding under the carpet for the optical range has been solved with the use of metama terials in the form of dielectrics structured on a sub wavelength scale [140142]. The device operating in the wavelength range 1.41.8 m was fabricated from the low loss perforated silicon structure [140]. The composite metamaterial consisting of silicon and sili con dioxide was used to fabricate the invisibility device for a wavelength of 1.55 m [141]. One more device operating in the range of a wavelength of 1.5 m was designed on the basis of silicon nanofibers [142]. These results are prominent indications of advances made in the field of metamaterials. It should also be noted that one more strategy of hiding under the carpet that requires the use of a spa tially homogeneous anisotropic dielectric material for the implementation was theoretically investigated in [143]. In particular, it was demonstrated that the two dimensional variant of this invisibility device operat ing in the IR range (13722000 nm) can be fabricated from layered structures consisting of silicon layers sep arated by air gaps.
2010

546

ZHILIN, SHEPILOV

METAMATERIALS AND TRANSFORMATION OPTICS The methods for transforming coordinates that were used for constructing the theoretical models of invisibility cloaks [90, 91] and, especially, the method employed in [91] have attracted considerable attention as a tool for solving a number of problems in the field of electrodynamics and optics and have received the special name, i.e., the transformation optics. The application of these methods leads to consideration of media with unusual electromagnetic properties. These media can be prepared, if it is possible, only through artificial controlled structuring of materials with the use of calculations. Therefore, the capabilities of transformation optics are closely related to progress in the field of theoretical development and fabrication of metamaterials. Let us briefly review the works on transformation optics. Leonhardt and Philbin [144] demonstrated that the general theory of relativity provides a theoret ical tool for developing devices from metamaterials. The employment of the covariant form of the Max wells equations, which is invariant with respect to the general spacetime transformations, allowed the authors to describe not only the invisibility devices but also a perfect lens from a negative refractive index material with the use of space transformations and the optical AharonovBohm effect and artificial black holes with the use of general spacetime transforma tions. Schurig et al. [145] also used the covariant form of the Maxwells equations. It was noted that the formu las derived for the transformation of the permittivities and permeabilities in space transformations are equiv alent to those obtained in [113]. The authors derived the expressions for the permittivity and permeability tensors in the Cartesian coordinate system for spheri cal and cylindrical cloaks that correspond to the linear transformation of the radius (20) in contrast to rela tionships (21) and (23) in the spherical and cylindrical coordinates obtained in [91] and [121], respectively. It is the authors opinion [145] that the Cartesian coor dinate system is more convenient for determining the permittivity and permeability in cloaks with other geometries and for integrating the ray equations in the transformed space. For the aforementioned cloaks of two types, the ray equations were obtained in [145] and the results of their integration are presented in figures. More recently, the space transformation method was used for theoretically designing media rotating electromagnetic fields [146] and optical elements of different purposes [147], including those magnifying without geometric aberrations. In [148150, 129], the method was applied in theoretical consideration of the possibility of fabricating invisibility cloaks for a fre quency range. In 2008, a number of works on transformation optics increased significantly. We restrict ourselves

only to mention of some of them. A brief excursus in this field of investigations is given in [151]. A detailed analysis of the principles of transformation optics and a number of its applications is presented in the review [152]. Invisibility cloaks of different shapes were con sidered in [153160], and a radiation concentrator was also examined in [158]. Quasi perfect invisibility cloaks with windows for the exchange of information and matter with the environment (open cloaks) were theoretically studied in [161]. The transformation optics was used to develop the techniques for cloaking objects on a dielectric surface [162] and, as was noted above, under a reflecting (metallic) surface [138]; in this case, devices of the last type were fabricated from metamaterials [139142], including those for the optical spectral range [140142]. Several reflection less optical devices characterized by an unusual elec tromagnetic behavior [163], beam bends, beam expanders [164], beam splitters, and other devices [165] were designed in the framework of transforma tion optics. On the basis of transformation optics, Tyc and Leonhardt [166] proposed the method that pro vides a way of weakening the singularity of the refrac tive index due to the changeover to an anisotropic medium and can be used as a tool for implementing some proposed optical devices or for extending their operating range of frequencies. Optical media based on metamaterials for the rotation of electromagnetic wave fronts in the microwave spectral range (10 GHz) were theoretically developed and experi mentally implemented [167]. Kundtz et al. [168] noted that the transformations of a passive space, i.e., the space that does not contain sources (currents, charges) were previously considered in the literature. Examples of transformations in the presence of sources were examined in [168]. The authors believed that the proposed approach also holds promise for designing antennas. The case of sin gular sources (surface and linear) was considered in [169]. It should also be noted that the idea of a super antenna based on the representation of the three dimensional space as a set of sheets with the subse quent performance of two dimensional conformal transformations on each sheet was discussed in [170]. The idea underlying the transformation optics can be considered from more general positions as invari ance of equations that describe particular physical processes with respect to space transformations. In this extended interpretation, it can be used, for exam ple, in acoustics, electrostatics, and quantum mechanics [171]. Correspondingly, apart from trans formation optics, transformation acoustics [172] was developed and acoustic metamaterials with a negative refractive index of sound [173, 174], acoustic super lenses [175], and metamaterials for the solution of problems of acoustic cloaking [176] were considered theoretically. There are works on the experimental implementation of devices based on acoustic metama terials (see, for example, [177, 178]); however, their
Vol. 36 No. 5 2010

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE

547

number is considerably smaller than the number of theoretical papers. On the whole, the number of works in this field is rather large; however, their discussion is beyond the scope of the present review. CONCLUSIONS The idea of fabrication of negative refractive index media, which were theoretically considered by Vese lago in 1967, led to the development of artificially structured materials, i.e., metamaterials, and to the formation of a new direction in materials science. An important impetus to the design of metamaterials was given by the Pendry idea (put forward in 2000) regard ing the creation of a perfect lens (superlens) on their basis. The two dimensional metamaterial with a nega tive refractive index of 2.7, which was determined directly from refraction experiments, for radiation with a wavelength of approximately 3 cm was first fab ricated in 2001. The investigations performed in sub sequent years made it possible to progress in the range of higher frequencies. A bulk nanostructured metama terial in the form of a multilayer metaldielectric fish net structure was fabricated in 2008. The refractive index of this metamaterial for radiation propagating perpendicular to the layer plane was determined directly from refraction experiments and amounted, for example, to 1.2 at a wavelength of 1.8 m. All the metamaterials of this type prepared to date are sub stantially anisotropic and characterized by a high absorption, because their structure contains metal structural elements. Isotropic metamaterials with a negative refraction index have not been fabricated to date, even though hypothetical methods for their preparation have been proposed in the literature. A decrease in the losses is a fundamentally important problem of metamaterials science. Investigations in the field of metamaterials have revealed that their electric and magnetic responses to electromagnetic radiation can be controlled over a wide range. In 20052006, this allowed one to raise the question as to the possibility of producing two types of invisible objects: objects that have sizes smaller than or of the order of the wavelength (first type) and invisible cloaks (second type) that conceal the objects contained in them from an external observer. Some aspects of this question were discussed in earlier works, which, however, at that time, did not attract wide attention and became significant only after the advent of metamaterials. In particular, in 1961, Dolin very closely came to the theoretically solution to the problem of objects of the second type. By using the space transformation method, he pre dicted the possibility of forming invisible optical inho mogeneities and noted a key role of the anisotropy of materials in the creation of invisibility. In 2006 and later, the space transformation method had been used to theoretically design invisible cloaks and other devices and underlain a new direction in optics, i.e.,
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

transformation optics. The prospects of experimental fabrication of invisible cloaks with the theoretically predicted design and the transformation optics have given a new impetus to the development of metamate rials. For example, the fabrication of invisibility cloaks requires anisotropic materials for which the permittiv ity and permeability are specified functions of the coordinates. A metamaterial of this type was prepared when implementing the simplified invisibility cloak that operates in the gigahertz range of frequencies and noticeably reduces the observability of the cloaked object. Spatially inhomogeneous metamaterials based on nanostructured silicon were designed in order to demonstrate interesting devices that operate accord ing to the principle of hiding under the carpet. Despite considerable progress, the invisibility problem is far from solution. Since the preparation of metamaterials is an expensive process, the parameters of structural ele ments cannot be chosen by the trial and error method. These parameters have been determined using modern tools of numerical simulation, without which the design of new metamaterials is impossible. Metamaterials for the optical spectral range cannot be fabricated without using nanotechnologies. The idea of metamaterials is important not only for optics and electromagnetism but also for acoustics. ACKNOWLEDGMENTS This study was supported by the Russian Founda tion for Basic Research (project no. 10 03 00458 a). REFERENCES
1. Born, M. and Wolf, E., Principles of Optics: Electro magnetic Theory of Propagation, Interference, and Dif fraction of Light, London: Pergamon, 1959. Translated under the title Osnovy optiki, Moscow: Nauka, 1973. 2. Landau, L.D. and Lifshitz, E.M., Teoreticheskaya fizika, Volume 8: Elektrodinamika sploshnykh sred, Moscow: Nauka, 1982. Translated under the title Course of Theoretical Physics, Volume 8: Electrody namics of Continuous Media, Oxford: Butterworth Heinemann, 1984. 3. Veselago, V.T., The Electrodynamics of Substances with Simultaneously Negative Values of and , Usp. Fiz. Nauk, 1967, vol. 92, no. 3, pp. 517526 [Sov. Phys. Usp. (Engl. transl.), 1968, vol. 10, no. 4, pp. 509514]. 4. Mandelshtam, L.I., Lectures on Certain Problems of the Theory of Oscillations, in Polnoe sobranie trudov (Complete Collection of Works), Moscow: Academy of Sciences of the Soviet Union, 1950, vol. 5, pp. 428 467 [in Russian]. 5. Lamb, H., On Group Velocity, Proc. London Math. Soc., 1904, vol. 1, pp. 473479. 6. Schuster, A., An Introduction to the Theory of Optics, London: Edward Arnold, 1904.
2010

548

ZHILIN, SHEPILOV 24. Tretyakov, S.A., Research on Negative Refraction and Backward Wave Media: A Historical Perspective, in Collection of Papers of EPFL Latsis Symposium 2005 Negative Refraction: Revisiting Electromagnetics from Microwaves to Optics, Lausanne, Switzerland, 2005, pp. 3035. 25. Veselago, V., Braginsky, L., Shklover, V., and Hafner, C., Negative Refraction Index Materials, J. Comput. Theor. Nanosci., 2006, vol. 3, no. 2, pp. 130. 26. Kotthaus, J.P. and Jaccarino, V., Antiferromagnetic Resonance Linewidths in MnF2, Phys. Rev. Lett., 1972, vol. 28, no. 25, pp. 16491652. 27. Grunberg, P. and Metawe, F., Light Scattering from Bulk and Surface Spin Waves in EuO, Phys. Rev. Lett., 1977, vol. 39, no. 24, pp. 15611565. 28. Sandercock, J.R. and Wettling, W., Light Scattering from Surface and Bulk Thermal Magnons in Iron and Nickel, J. Appl. Phys., 1979, vol. 50, no. B11, pp. 77847789. 29. Camley, R.E. and Mills, D.L., Surface Polaritons on Uniaxial Antiferromagnets, Phys. Rev. B: Condens. Matter, 1982, vol. 26, no. 3, pp. 12801287. 30. Remer, L., Luthi, B., Sauer, H., Geick, R., and Cam ley, R.E., Nonreciprocal Optical Reflection of the Uniaxial Antiferromagnet MnF2, Phys. Rev. Lett., 1986, vol. 56, no. 25, pp. 27522754. 31. Dumelow, T., Camley, R.E., Abraha, K., and Tilley, D.R., Nonreciprocal Phase Behavior in Reflection of Elec tromagnetic Waves from Magnetic Materials, Phys. Rev. B: Condens. Matter, 1998, vol. 58, no 2, pp. 897 908. 32. Pendry, J.B. and OBrien, S., Very Low Frequency Plasma, J. Phys.: Condens. Matter, 2002, vol. 14, no. 32, pp. 74097416. 33. OBrien, S., MacPeake, D., Ramakrishna, S.A., and Pendry, J.B., Near Infrared Photonic Band Gaps and Nonlinear Effects in Negative Magnetic Metamaterials, Phys. Rev. B: Condens. Matter, 2004, vol. 69, no. 24, arti cle 241101 (4 pages). 34. Wiltshire, M.C.K., Pendry, J.B., Young, I.R, Lark man, D.J., Gilderdale, D.J., and Hajnal, J.V., Micro structured Magnetic Materials for RF Flux Guides in Magnetic Resonance Imaging, Science (Washington), 2001, vol. 291, no. 5505, pp. 848851. 35. Wiltshire, M.C.K, Hajnal, J.V., Pendry, J.B., Edwards, D.J., and Stevens, C.J., Metamaterial Endoscope for Magnetic Field Transfer: Near Field Imaging with Magnetic Wires, Opt. Express, 2003, vol. 11, no. 7, pp. 709715. 36. Pendry, J.B., Holden, A.J., Robbins, D.J., and Stewart, W.J., Magnetism from Conductors and Enhanced Nonlinear Phenomena, IEEE Trans. Microwave Theor. Tech., 1999, vol. 47, no. 11, pp. 20752084. 37. Smith, D.R., Padilla, W.J., Vier, D.C., Nemat Nasser, S.C., and Schultz, S., Composite Medium with Simultaneously Negative Permeability and Per mittivity, Phys. Rev. Lett., 2000, vol. 84, no. 18, pp. 41844187. 38. Yen, T.J., Padilla, W.J., Fang, N., Vier, D.C., Smith, D.R., Pendry, J.B., Basov, D.N., and Zhang, X., Terahertz Magnetic Response from Artificial Materials, Science
Vol. 36 No. 5 2010

7. Mandelshtam, L.I., Group Velocity in the Crystal Lat tice, Zh. Eksp. Teor. Fiz., 1945, vol. 15, no. 9, pp. 475 478. 8. Ramakrishna, S.A., Physics of Negative Refractive Index Materials, Rep. Prog. Phys., 2005, vol. 68, no. 2, pp. 449521. 9. Bliokh, K.Yu. and Bliokh, Yu.K., What Are the Left Handed Media and What Is Interesting about Them? Usp. Fiz. Nauk, 2004, vol. 174, no. 4, pp. 439447 [Phys.Usp. (Engl. transl.), 2004, vol. 47, no. 4, pp. 393400]. 10. Pendry, J.B., Negative Refraction Makes a Perfect Lens, Phys. Rev. Lett., 2000, vol. 85, no. 18, pp. 3966 3969. 11. Hooft, G.W.'t., Comment on Negative Refraction Makes a Perfect Lens, Phys. Rev. Lett., 2001, vol. 87, no. 24, article 249701 (1 page). 12. Williams, J.M., Some Problems with Negative Refrac tion, Phys. Rev. Lett., 2001, vol. 87, no. 24, article 249703 (1 page). 13. Garcia, N. and Nieto Vesperinas, M., Left Handed Materials Do not Make a Perfect Lens, Phys. Rev. Lett., 2002, vol. 88, no. 20, article 207403 (4 pages). 14. Pendry, J., Replay, Phys. Rev. Lett., 2001, vol. 87, no. 24, article 249 702 (1 page). 15. Pendry, J., Replay, Phys. Rev. Lett., 2001, vol. 87, no. 24, article 249 704 (1 page). 16. Shen, J.T. and Platzman, P.M., Near Field Imaging with Negative Dielectric Constant Lenses, Appl. Phys. Lett., 2002, vol. 80, no. 18, pp. 32863288. 17. Smith, D.R., Schurig, D., Rosenbluth, M., Schultz, S., Ramakrishna, S.A., and Pendry, J.B., Limitations on Subdiffraction Imaging with a Negative Refractive Index Slab, Appl. Phys. Lett., 2003, vol. 82, no. 10, pp. 15061508. 18. Kolinko, P. and Smith, D., Numerical Study of Elec tromagnetic Waves Interacting with Negative Index Materials, Opt. Express, 2003, vol. 11, no. 7, pp. 640 648. 19. Cummer, S.A., Simulated Causal Subwavelength Focusing by a Negative Refractive Index Slab, Appl. Phys. Lett., 2003, vol. 82, no. 10, pp. 15031505. 20. Rao, X.S. and Ong, C.K., Subwavelength Imaging by a Left Handed Material Superlens, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2003, vol. 68, no. 6, arti cle 067 601 (3 pages). 21. Huang, X. and Zhou, L., Modulating Image Oscilla tions in Focusing by a Metamaterial Lens: Time Dependent Greens Function Approach, Phys. Rev. B: Condens. Matter, 2006, vol. 74, no. 4, article 045 123 (8 pages). 22. Cubukcu, E., Aydin, K., Ozbay, E., Foteinopolou, S., and Soukoulis, C.M., Subwavelength Resolution in a Two Dimensional Photonic Crystal Based Super lens, Phys. Rev. Lett., 2003, vol. 91, no. 20, article 207401 (4 pages). 23. Grbic, A. and Eleftheriades, G.V., Overcoming the Diffraction Limit with a Planar Left Handed Trans mission Line Lens, Phys. Rev. Lett., 2004, vol. 92, no. 11, article 117403 (4 pages).

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE (Washington), 2004, vol. 303, no. 5663, pp. 1494 1496. 39. Linden, S., Enkrich, C., Wegener, M., Zhou, J., Koschny, T., and Soukoulis, C.M., Magnetic Response of Metamaterials at 100 Terahertz, Science (Washington), 2004, vol. 306, no. 5700, pp. 1351 1353. 40. Cai, W., Chettiar, U.K., Kildishev, A.V., and Shalaev, V.M., Optical Cloaking with Metamaterials, Nat. Photonics, 2004, vol. 1, pp. 224227. 41. Podolskiy, V.A., Sarychev, A.K., and Shalaev, V.M., Plasmon Modes in Metal Nanowires and Left Handed Materials, J. Nonlinear Opt. Phys. Mater., 2002, vol. 11, no. 1, pp. 6574. 42. Kildishev, A.V., Cai, W., Chettiar, U.K., Yuan, H.K., Sarychev, A.K., Drachev, V.P., and Shalaev, V.M., Negative Refraction Index in Optics of MetalDielec tric Composites, J. Opt. Soc. Am. B, 2006, vol. 23, no. 3, pp. 423433. 43. Pendry, J.B., Holden, A.J., Stewart, W.J., and Youngs, I., Extremely Low Frequency Plasmons in Metallic Mesostructures, Phys. Rev. Lett., 1996, vol. 76, no. 25, pp. 47734776. 44. Pendry, J.B., Holden, A.J., Robbins, D.J., and Stew ard, W.J., Low Frequency Plasmons in Thin Wire Structures, J. Phys.: Condens. Matter, 1998, vol. 10, no. 22, pp. 47854809. 45. Raether, H., Excitation of Plasmons and Interband Transitions by Electrons, Berlin: Springer, 1980. 46. Raether, H., Surface Plasmons on Smooth and Rough Surfaces and on Gratings, Berlin: Springer, 1988. 47. Wu, D., Fang, K., Sun, C., Zhang, X., Padilla, W.J., Basov, D.K., Smith, D.R., and Schultz, S., Terahertz Plasmonic High Pass Filter, Appl. Phys. Lett., 2003, vol. 83, no. 1, pp. 201203. 48. Marqus, R., Medina, F., and Rafli El Idrissi, R., Role of Bianisotropy in Negative Permeability and Left Handed Metamaterials, Phys. Rev. B: Condens. Matter, 2002, vol. 65, no. 14, article 144440 (6 pages). 49. Gay Balmaz, P. and Martin, O.J.F., Electromagnetic Resonances in Individual and Coupled Split Ring Resonators, J. Appl. Phys., 2002, vol. 92, no. 5, pp. 29292936. 50. Shelby, R.A., Smith, D.R., and Schultz, S., Experi mental Verification of a Negative Index of Refraction, Science (Washington), 2001, vol. 292, no. 5514, pp. 7779. 51. Parazzoli, C.G., Greegor, R., Li, K., Kontenbah, B.E.C., and Tanielian, M.H., Experimental Verification and Simulation of Negative Index of Refraction Using Snells Law, Phys. Rev. Lett., 2003, vol. 90, no. 10, arti cle 107401 (4 pages). 52. Li, K., McLean, S.J., Greegor, R.B., Parazzoli, C.G., and Tanielian, M.H., Free Space Focused Beam Characterization of Left Handed Materials, Appl. Phys. Lett., 2003, vol. 82, no. 15, pp. 25352537. 53. Gokkavas, M., Guven, K., Bulu, L., Aydin, K., Pen ciu, R.S., Kafesaki, M., Soukoulis, C.M., and Ozbay, E., Experimental Demonstration of a Left Handed Metamaterial Operating at 100 GHz, Phys. Rev. B: Condens. Matter, 2006, vol. 73, no. 19, article 193104 (4 pages).
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

549

54. Hsu, A. C., Cheng, Y. K., Chen, K. K., Chern, J. L., Wu, S. C., Chen, C. F., Chang, H., Lien, Y. K., and Shy, J. T., Far Infrared Resonance in Split Ring Res onators, Jpn. J. Appl. Phys., 2004, vol. 43, no. 2A, pp. L176L179. 55. Enkrich, C., Wegener, M., Linden, S., Burger, S., Zschiedrich, L., Schmidt, F., Zhou, J.F., Koschny, T., and Soukoulis, C.M., Magnetic Metamaterials at Telecommunication and Visible Frequencies, Phys. Rev. Lett., 2005, vol. 95, no. 20, article 203901 (4 pages). 56. Moser, H.O., Casse, B.D.F., Wilhelmi, O., and Saw, B.T., Terahertz Response of a Microfabricated Rod Split Ring Resonator Electromagnetic Metama terial, Phys. Rev. Lett., 2005, vol. 94, no. 6, article 063901 (4 pages). 57. Klar, T.A., Kildishev, A.V., Drachev, V.P., and Shalaev, V.M., Negative Index Metamaterials: Going Optical, IEEE J. Selected Top. Quantum Electron., 2006, vol. 12, no. 6, pp. 11061115. 58. Lagarkov, A.N. and Sarychev, A.K., Electromagnetic Properties of Composites Containing Elongated Con ducting Inclusions, Phys. Rev. B: Condens. Matter, 1996, vol. 53, no. 10, pp. 63186336. 59. Podolskiy, V.A., Sarychev, A.K., and Shalaev, V.M., Plasmon Modes in Metal Nanowires and Left Handed Materials, J. Nonlinear Opt. Phys. Mater., 2002, vol. 11, no. 1, pp. 6574. 60. Dolling, G., Enkrich, C., Wegener, M., Zhou, J.F., Soukoulis, C.M., and Linden, S., Cut Wire Pairs and Plate Pairs as Magnetic Atoms for Optical Metamate rials, Opt. Lett., 2005, vol. 30, no. 23, pp. 31983200. 61. Panina, L.V., Grigorenko, A.N., and Makhnovskiy, D.P., Optomagnetic Composite Medium with Conducting Nanoelements, Phys. Rev. B: Condens. Matter, 2002, vol. 66, no. 15, article 155411 (17 pages). 62. Podolskiy, V.A., Sarychev, A.K., and Shalaev, V.M., Plasmon Modes in Metal Nanowires and Left Handed Materials, Opt. Express, 2003, vol. 11, no. 7, pp. 735745. 63. Podolskiy, V.A., Sarychev, A.K., Narimanov, E.E., and Shalaev, V.M., Resonant Light Interaction with Plas monic Nanowire Systems, J. Opt. A: Pure Appl. Opt., 2005, vol. 7, pp. S32S37. 64. Shalaev, V.M., Cai, W., Chettiar, U.K., Yuan, K.K., Sarychev, A.K., Drachev, V.P., and Kildishev, A.V., Negative Index of Refraction in Optical Metamaterials, Opt. Lett., 2005, vol. 30, no. 24, pp. 33563358. 65. Drachev, V.P., Cai, W., Chettiar, U.K., Yuan, K.K., Sarychev, A.K., Kildishev, A.V., Klimec, G., and Shalaev, V.M., Experimental Verification of an Optical Negative Index Material, Laser Phys. Lett., 2006, vol. 3, no. 1, pp. 4955. 66. Smith, D.R., Schultz, S., Marko s , P., and Soukou lis, C.M., Determination of Effective Permittivity and Permeability of Metamaterials from Reflection and Transmission Coefficients, Phys. Rev. B: Condens. Matter, 2002, vol. 65, no. 19, article 195104 (5 pages). 67. Grigorenko, A.N., Geim, A.K., Gleeson, H.F., Zhang, Y., Firsov, A.A., Khrushchev, I.Y., and Petro vic, J., Nanofabricated Media with Negative Perme
2010

550

ZHILIN, SHEPILOV 82. Kussow, A. G., Akyurtlu, A., and Angkawisittpan, N., Optically Isotropic Negative Index of Refraction Metamaterial, Phys. Status Solidi B, 2008, vol. 245, no. 5, pp. 992997. 83. Bohren, C.F. and Huffman, D.R., Absorption and Scattering of Light by Small Particles, New York: Wiley, 1983. Translated under the title Pogloshchenie i rassey anie sveta malymi chastitsami, Moscow: Mir, 1986. 84. Yannopapas, V. and Moroz, A., Negative Refractive Index Metamaterials from Inherently Nonmagnetic Materials for Deep Infrared to Terahertz Frequency Ranges, J. Phys.: Condens. Matter, 2005, vol. 17, no. 25, pp. 37173734. 85. Boltasseva, A. and Shalaev, V.M., Fabrication of Opti cal Negative Index Metamaterials: Recent Advances and Outlook, Metamaterials, 2008, vol. 2, no. 1, pp. 117. 86. Smolyaninov, I.I., Smolyaninova, V.N., Kildishev, A.V., and Shalaev, V.M., Anisotropic Metamaterials Emu lated by Tapered Waveguides: Application to Optical Cloaking, Phys. Rev. Lett., 2009, vol. 102, no. 21, arti cle 213901 (4 pages). 87. Fante, R.L. and McCormack, M.T., Reflection Proper ties of the Salisbury Screen, IEEE Trans. Antenn. Propa gation, 1988, vol. 36, no. 10, pp. 14431454. 88. Ward, J., Towards Invisible Glass, Vacuum, 1972, vol. 22, no. 9, pp. 369375. 89. Al, A. and Engheta, N., Achieving Transparency with Plasmonic and Metamaterial Coatings, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2005, vol. 72, no. 1, article 016 623 (9 pages). 90. Leonhardt, U., Optical Conformal Mapping, Science (Washington), 2006, vol. 312, no. 5781, pp. 1777 1780. 91. Pendry, J.B., Schurig, D., and Smith, D.R., Control ling Electromagnetic Fields, Science (Washington), 2006, vol. 312, no. 5781, pp. 17801782. 92. Andreev, N.S., Scattering of Visible Light by Glasses Undergoing Phase Separation and Homogenization, J. Non Cryst. Solids, 1978, vol. 30, no. 2, pp. 99126. 93. Walker, C.B. and Guinier, A., An X Ray Investigation of Age Hardening in Alag, Acta Metall., 1953, vol. 1, no 5, pp. 568577. 94. Shatilov, A.V., Anomalous Scattering as the Case of Scattering from a System of Particles, Opt. Spektrosk., 1962, vol. 13, no. 5, pp. 728733. 95. Goldstein, M., Theory of Scattering for Diffusion Controlled Phase Separation, J. Appl. Phys., 1963, vol. 34, no. 7, pp. 19281934. 96. Kerker, M., Invisible Bodies, J. Opt. Soc. Am., 1975, vol. 65, no. 4, pp. 376379. 97. Chew, H. and Kerker, M., Abnormally Low Electro magnetic Scattering Cross Sections, J. Opt. Soc. Am., 1976, vol. 66, no. 5, pp. 445449. 98. Kerker, M. and Blatchford, C.G., Elastic Scattering, Absorption, and Surface Enhanced Raman Scattering by Concentric Spheres Comprised of a Metallic and a Dielectric Region, Phys. Rev. B: Condens. Matter, 1982, vol. 26, no. 8, pp. 40524063. 99. Aden, A.L. and Kerker, M., Scattering of Electromag netic Waves from Two Concentric Spheres, J. Appl. Phys., 1951, vol. 22, no. 10, pp. 12421246.
Vol. 36 No. 5 2010

ability at Visible Frequencies, Nature (London), 2005, vol. 438, no. 7066, pp. 335338. 68. Zhang, S., Fan, W., Malloy, K.J., Brueck, S.R.J., Panoiu, N.C, and Osgood, R.M., Near Infrared Dou ble Negative Metamaterials, Opt. Express, 2005, vol. 13, no. 13, pp. 49224930. 69. Zhang, S., Fan, W., Panoiu, N.C., Malloy, K.J., Osgood, R.M., and Brueck, S.R.J., Experimental Demonstration of Near Infrared Negative Index Materials, Phys. Rev. Lett., 2005, vol. 95, no. 13, article 137404 (4 pages). 70. Zhang, S., Fan, W., Malloy, K.J., Brueck, S.R.J., Panoiu, N.C., and Osgood, R.M., Demonstration of MetalDielectric Negative Index Metamaterials with Improved Performance at Optical Frequencies, J. Opt. Soc. Am. B, 2006, vol. 23, no. 3, pp. 434438. 71. Falcone, F., Lopetegi, T., Laso, M.A.G., Baena, J.D., Bonache, J., Beruete, M., Marques, R., Martin, F., and Sorolla, M., Babinet Principle Applied to the Design of Metasurfaces and Metamaterials, Phys. Rev. Lett., 2004, vol. 93, no. 19, article 197401 (4 pages). 72. Dolling, G., Enkrich, C., Wegener, M., Soukoulis, C.M., and Linden, S., Low Loss Negative Index Metamate rial at Telecommunication Wavelengths, Opt. Lett., 2006, vol. 31, no. 12, pp. 18001802. 73. Dolling, G., Wegener, M., Soukoulis, C.M., and Lin den, S., Negative Index Metamaterial at 780 nm Wavelength, Opt. Lett., 2007, vol. 32, no. 1, pp. 5355. 74. Dolling, G., Enkrich, C., Wegener, M., Soukoulis, C.M., and Linden, S., Simultaneous Negative Phase and Group Velocity of Light in a Metamaterial, Science (London), 2006, vol. 312, no. 5775, pp. 892894. 75. Zhang, S., Fan, W., Panoiu, N.C., Malloy, K.J., Osgood, R.M., and Brueck, S.R.J., Optical Negative Index Bulk Metamaterials Consisting of 2D Perforated MetalDielectric Stacks, Opt. Express, 2006, vol. 14, no. 15, pp. 67786787. 76. Valentine, J., Zhang, S., Zentgraf, T., Ulin Avila, E., Genov, D.A., Bartal, G., and Zhang, X., Three Dimensional Optical Metamaterial with a Negative Refractive Index, Nature (London), 2008, vol. 455, no. 7211, pp. 376379. 77. Navarro Cia, M., Beruete, M., Sorolla, M., and Campillo, I., Negative Refraction in a Prism Made of Stacked Subwavelength Hole Arrays, Opt. Express, 2008, vol. 16, no. 2, pp. 560566. 78. Lezec, H.J., Dionne, J.A., and Atwater, H.A., Nega tive Refraction at Visible Frequencies, Science (Wash ington), 2007, vol. 316, no. 5823, pp. 430432. 79. Koschny, T., Zhang, L., and Soukoulis, C.M., Isotro pic Three Dimensional Left Handed Metamaterials, Phys. Rev. B: Condens. Matter, 2005, vol. 71, no. 12, article 121103 (4 pages). 80. Andryieuski, A., Malureanu, R., and Lavrinenko, A., Nested Structures Approach in Designing an Isotropic Negative Index Material for Infrared, J. Eur. Opt. Soc., Rapid Publ., 2009, vol. 4, article 09 003 (7 pages). 81. Kussow, A. G., Akyurtlu, A., Semichaevsky, A., and Angkawisittpan, N., MgB2 Based Negative Refraction Index Metamaterial at Visible Frequencies: Theoreti cal Analysis, Phys. Rev. B: Condens. Matter, 2007, vol. 76, no. 19, article 195123 (7 pages).

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE 100. Al, A. and Engheta, N., Polarizabilities and Effective Parameters for Collections of Spherical Nanoparticles Formed by Pairs of Concentric Double Negative, Sin gle Negative, and/or Double Positive Metamaterial Layers, J. Appl. Phys., 2005, vol. 97, no. 9, article 094310 (12 pages). 101. Al, A. and Engheta, N., Erratum: Polarizabilities and Effective Parameters for Collections of Spherical Nanoparticles Formed by Pairs of Concentric Double Negative, Single Negative, and/or Double Positive Metamaterial Layers [J. Appl. Phys., 2005, vol. 97, 094310], J. Appl. Phys., 2006, vol. 99, no. 6, article, 069901(E) (1 page). 102. Zhou, X. and Hu, G,. Design for Electromagnetic Wave Transparency with Metamaterials, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2006, vol. 74, no. 1, article 026607 (8 pages). 103. Maxwell Garnett, J.C., Colours in Metal Glasses and in Metallic Films, Philos. Trans. R. Soc., Ser. A, 1904, vol. 203, pp. 385420. 104. Zhou, X. and Hu, G., Linear and Nonlinear Dielectric Properties of Particulate Composites at Finite Con centration, Appl. Math. Mech., 2006, vol. 27, no. 8, pp. 10211030. 105. Al, A. and Engheta, N., Plasmonic Materials in Transparency and Cloaking Problems: Mechanism, Robustness, and Physical Insights, Opt. Express, 2007, vol. 15, no. 6, pp. 33183332. 106. Al, A. and Engheta, N., Cloaking and Transparency for Collection of Particles with Metamaterial and Plas monic Covers, Opt. Express, 2007, vol. 15, no. 12, pp. 75787590. 107. Al, A. and Engheta, N., Multifrequency Optical Invisibility Cloak with Layered Plasmonic Shells, Phys. Rev. Lett., 2008, vol. 100, no. 11, article 113901 (4 pages). 108. Silveirinha, M.G., Al, A., and Engheta, N., Parallel Plate Metamaterials for Cloaking Structures, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2007, vol. 75, no. 3, article 036603 (16 pages). 109. Nachman, A.I., Reconstructions from Boundary Measurements, Ann. Math., Ser. 2, 1988, vol. 128, no. 3, pp. 531576. 110. Wolf, E. and Habashy, T., Invisible Bodies and Uniqueness of the Inverse Scattering Problem, J. Mod. Opt., 1993, vol. 40, no. 5, pp. 785792. 111. Leonhardt, U., Notes on Conformal Invisibility Devices, New J. Phys., 2006, vol. 8, article 118 (16 pages). 112. Ochiai, T., Leonhardt, U., Nacher, J.C., A Novel Design of Dielectric Perfect Invisibility Devices, J. Math. Phys., 2008, vol. 49, no. 3, article 032903 (13 pages). 113. Ward, A.J. and Pendry, J.B., Refraction and Geometry in Maxwells Equations, J. Mod. Opt., 1996, vol. 43, no. 4, pp. 773793. 114. Korn, G. and Korn, T., Mathematical Handbook for Scientists and Engineers, New York: McGraw Hill, 1961. Translated under the title Spravochnik po matema tike, Moscow: Nauka, 1978, pp. 186188, Table 6.5 1. 115. Smith, D.R., Vier, D.C, Koschny, Th., and Soukou lis, C.M., Electromagnetic Parameter Retrieval from
GLASS PHYSICS AND CHEMISTRY Vol. 36 No. 5

551

Inhomogeneous Metamaterials, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2005, vol. 71, no. 3, arti cle 036617 (11 pages). 116. Driscoll, T., Basov, D.N, Starr, A.F., Rye, P.M., Nemat Nasser, S., Schurig, D., and Smith, D.R., Free Space Microwave Focusing by a Negative Index Gradient Lens, Appl. Phys. Lett., 2006, vol. 88, no. 8, article 081101 (3 pages). 117. Greegor, R.B., Parazzoli, C.G., Nielsen, J.A., Thompson, M.A., Tanielian, M.H., and Smith, D.R., Simulation and Testing of a Graded Negative Index of Refraction Lens, Appl. Phys. Lett., 2005, vol. 87, no. 9, article 091114 (3 pages). 118. Schurig, D., Mock, J.J., Justice, B.J., Cummer, S.A., Pendry, J.B., Starr, A.F., and Smith, D.R., Metamate rial Electromagnetic Cloak at Microwave Frequen cies, Science (Washington), 2006, vol. 314, no. 5801, pp. 977980. 119. Rozanov, N.N., Invisibility: For and against, Priroda (Moscow), 2008, no. 6; http://elementy.ru/lib/430669. 120. Dolin, L.S., On the Possibility of Comparing Three Dimensional Electromagnetic Systems with Inhomo geneous Anisotropic Filling, Izv. Vyssh. Uchebn. Zaved., Radiofiz., 1961, vol. 4, no. 5, pp. 964967. 121. Cummer, S.A., Popa, B. L., Schurig, D., Smith, D.R., and Pendry, J., Full Wave Simulations of Electromag netic Cloaking Structures, Phys. Rev. E: Stat., Nonlin ear, Soft Matter Phys., 2006, vol. 74, no. 3, article 036621 (5 pages). 122. Yan, M., Ruan, Z., and Qiu, M., Cylindrical Invisibil ity Cloak with Simplified Material Parameters Is Inherently Visible, Phys. Rev. Lett., 2007, vol. 99, no. 23, article 233901 (4 pages). 123. Yan, M., Ruan, Z., and Qiu, M., Scattering Charac teristics of Simplified Cylindrical Invisibility Cloaks, Opt. Express, 2007, vol. 15, no. 8, pp. 1777217782. 124. Chen, X., Grzegorczyk, T.M., Wu, B. L., Pacheco, J., and Kong, J.A., Robust Method to Retrieve the Con stitutive Effective Parameters of Metamaterials, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2004, vol. 70, no. 1, article 016608 (7 pages). 125. Cai, W., Chettiar, U.K., Kildishev, A.V., Shalaev, V.M., and Milton, G.W., Nonmagnetic Cloak with Mini mized Scattering, Appl. Phys. Lett., 2007, vol. 91, no. 11, article 111105 (3 pages). 126. Weder, R., A Rigorous Analysis of High Order Elec tromagnetic Invisibility Cloaks, J. Phys. A: Math. Theor., 2008, vol. 41, no. 6, article 065207 (21 pages). 127. Cai, W., Chettiar, U.K., Kildishev, A.V., and Shalaev, V.M., Design for Optical Cloaking with High Order Trans formations, Opt. Express, 2008, vol. 16, no. 8, pp. 54445452. 128. Collins, P. and McGuirk, J., A Novel Methodology for Deriving Improved Material Parameter Sets for Sim plified Cylindrical Cloaks, J. Opt. A: Pure Appl. Opt., 2009, vol. 11, article 015104 (8 pages). 129. Leonhardt, U. and Tyc, T., Broadband Invisibility by Non Euclidean Cloaking, Science (Washington), 2009, vol. 323, no. 5910, pp. 110112. 130. Huang, Y., Feng, Y., and Jiang, T., Electromagnetic Cloaking by Layered Structure of Homogeneous Iso
2010

552

ZHILIN, SHEPILOV 146. Chen, H. and Chan, C.T., Transformation Media That Rotate Electromagnetic Fields, Appl. Phys. Lett., 2007, vol. 90, no. 24, article 241105 (3 pages). 147. Schurig, D., Pendry, J.B., and Smith, D.R., Transfor mation Designed Optical Elements, Opt. Express, 2007, vol. 15, no. 22, pp. 1477214782. 148. Chen, H., Liang, Z., Yao, P., Jiang, X., Ma, H., and Chan, C.T., Extending the Bandwidth of Electromag netic Cloaks, Phys. Rev. B: Condens. Matter, 2007, vol. 76, no. 24, article 241104 (4 pages). 149. Kildishev, A.V., Cai, W., Chettiar, U.K., and Shalaev, V.M., Transformation Optics: Approaching Broadband Electromagnetic Cloaking, New J. Phys., 2008, vol. 10, article 115029 (13 pages). 150. Farhat, M., Guenneau, S., Movchan, A.B., and Enoch, S., Achieving Invisibility over a Finite Range of Frequencies, Opt. Express, 2008, vol. 16, no. 8, pp. 56565661. 151. Shalaev, V.M., Transforming Light, Science (Washing ton), 2008, vol. 322, no. 5900, pp. 384386. 152. Leonhardt, U. and Philbin, T.G., Transformation Optics and the Geometry of Light, Prog. Opt., 2009, vol. 53, pp. 69152. 153. Yan, W., Yan, M., Ruan, Z., and Qiu, M., Coordinate Transformations Make Perfect Invisibility Cloaks with Arbitrary Shape, New J. Phys., 2008, vol. 10, article 043040 (13 pages). 154. You, Y., Kattawar, G.W., Zhai, P. W., and Yang, P., Invisibility Cloaks for Irregular Particles Using Coor dinate Transformations, Opt. Express, 2008, vol. 16, no. 9, pp. 61346145. 155. Jiang, W.X., Chin, J.Y., Li, Z., Cheng, Q., Liu, R., and Cui, T.J., Analytical Design of Conformally Invisible Cloaks for Arbitrarily Shaped Objects, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2008, vol. 77, no. 6, article 066 607 [6 pages). 156. Wang, W., Lin, L., Yang, X., Cui, J., Du, C., and Luo, X., Design of Oblate Cylindrical Perfect Lens Using Coordinate Transformation, Opt. Express, 2008, vol. 16, no. 11, pp. 80948105. 157. Nicolet, A., Zolla, F., and Guenneau, S., Electromag netic Analysis of Cylindrical Cloaks of an Arbitrary Cross Section, Opt. Lett., 2008, vol. 33, no. 14, pp. 15841586. 158. Rahm, M., Schurig, D., Roberts, D.A., Cummer, S.A., Smith, D.R., and Pendry, J.B., Design of Electromag netic Cloaks and Concentrators Using Form Invariant Coordinate Transformations of Maxwells Equations, Photonics Nanostruct., 2008, vol. 6, pp. 8795. 159. Li, C., Yao, K., and Li, F., Two Dimensional Electro magnetic Cloaks with Non Conformal Inner and Outer Boundaries, Opt. Express, 2008, vol. 16, no. 23, pp. 1936619374. 160. Wu, Q., Zhang, K., Meng, F., and Li, L. W., Material Parameters Characterization for Arbitrary N Sided Regular Polygonal Invisible Cloak, J. Phys. D: Appl. Phys., 2009, vol. 42, article 035408 (7 pages). 161. Ma, H., Qu, S., Xu, Z., and Wang, J., The Open Cloak, Appl. Phys. Lett., 2009, vol. 94, no. 10, article 103501 (3 pages).
Vol. 36 No. 5 2010

tropic Materials, Opt. Express, 2007, vol. 15, no. 18, pp. 1113311141. 131. Qiu, C. W., Hu, L., Xu, X., and Feng, Y., Spherical Cloaking with Homogeneous Isotropic Multilayered Structures, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2009, vol. 79, no. 4, article 047602 (4 pages). 132. Sun, J., Zhou, J., and Kang, L., Homogenous Isotro pic Invisible Cloak Based on Geometrical Optics, Opt. Express, 2008, vol. 16, no. 22, pp. 1776817773. 133. Lai, Y., Chen, H., Zhang, Z. Q., and Chan, C.T., Complementary Media Invisibility Cloak That Cloaks Objects at a Distance Outside the Cloaking Shell, Phys. Rev. Lett., 2009, vol. 102, no. 9, article 093901 (4 pages). 134. Kante, B., Germain, D., and de Lustrac, A., Experi mental Demonstration of a Nonmagnetic Metamate rial Cloak at Microwave Frequencies, Phys. Rev. B: Condens. Matter, 2009, vol. 80, no. 20, article 201104 (4 pages). 135. Liu, X., Li, C., Yao, K., Meng, X., Feng, W., Wu, B., and Li, F., Experimental Verification of Broadband Invisibility Using a Cloak Based on InductorCapac itor Networks, Appl. Phys. Lett., 2009, vol. 95, no. 19, article 191107 (3 pages). 136. Tretyakov, S., Alitalo, P., Luukkonen, O., and Simo vski, C., Broadband Electromagnetic Cloaking of Long Cylindrical Objects, Phys. Rev. Lett., 2009, vol. 103, no. 10, article 103905 (4 pages). 137. Smolyaninov, I.I., Hung, Y.J., and Davis, C.C., Two Dimensional Metamaterial Structure Exhibiting Reduced Visibility at 500 nm, Opt. Lett., 2008, vol. 33, no. 12, pp. 13421344. 138. Li, J. and Pendry, J.B., Hiding Under the Carpet: A New Strategy for Cloaking, Phys. Rev. Lett., 2008, vol. 101, no. 20, article 203901 (4 pages). 139. Liu, R., Ji, C., Mock, J.J., Chin, J.Y., Cui, T.J., and Smith, D.R., Broadband Ground Plane Cloak, Science (Washington), 2009, vol. 323, no. 5912, pp. 366369. 140. Valentine, J., Li, J., Zentgraf, T., Bartal, G., and Zhang, X., An Optical Cloak Made of Dielectrics, Nat. Mater., 2009, vol. 8, no. 7, pp. 568571. 141. Gabrielli, L.H., Cardenas, J., Poitras, C.B., and Lip son, M., Silicon Nanostructure Cloak Operating at Optical Frequencies, Nat. Photonics, 2009, vol. 3, no. 8, pp. 461463. 142. Lee, J.K., Blair, J., Tamma, V.A., Wu, Q., Rhee, S.J., Summers, C.J., and Park, W., Direct Visualization of Optical Frequency Invisibility Cloak Based on Silicon Nanorod Array, Opt. Express, 2009, vol. 17, no. 15, pp. 1292212928. 143. Xu, X., Feng, Y., Hao, Y., Zhao, J., and Jiang, T., Infrared Carpet Cloak Designed with Uniform Silicon Grating Structure, Appl. Phys. Lett., 2009, vol. 95, no. 18, article 184102 (3 pages). 144. Leonhardt, U. and Philbin, T.G., General Relativity in Electrical Engineering, New J. Phys., 2006, vol. 8, arti cle 247 (18 pages). 145. Schurig, D., Pendry, J.B., and Smith, D.R., Calcula tion of Material Properties and Ray Tracing in Trans formation Media, Opt. Express, 2006, vol. 14, no. 21, pp. 97949802.

GLASS PHYSICS AND CHEMISTRY

METAMATERIALS: A NEW DIRECTION IN MATERIALS SCIENCE 162. Zhang, P., Jin, Y., and He, S., Cloaking an Object on a Dielectric Half Space, Opt. Express, 2008, vol. 16, no. 5, pp. 31613166. 163. Rahm, M., Cummer, S.A., Schurig, D., Pendry, J.B., and Smith, D.R., Optical Design of Reflectionless Complex Media by Finite Embedded Coordinate Transformations, Phys. Rev. Lett., 2008, vol. 100, no. 6, article 063 903 (4 pages). 164. Rahm, M., Roberts, D.A., Pendry, J.B., and Smith, D.R., Transformation Optical Design of Adaptive Beam Bends and Beam Expanders, Opt. Express, 2008, vol. 16, no. 15, pp. 1155511567. 165. Zhai, T., Zhou, Y., Zhou, J., and Liu, D., Polarization Controller Based on Embedded Optical Transforma tion, Opt. Express, 2009, vol. 17, no. 20, pp. 17206 17213. 166. Tyc, T. and Leonhardt, U., Transmutation of Singular ities in Optical Instruments, New J. Phys., vol. 10, arti cle 115038 (8 pages). 167. Chen, H., Hou, V., Chen, S., Ao, X., Wen, W., and Chan, S.T., Design and Experimental Realization of a Broadband Transformation Media Field Rotator at Microwave Frequencies, Phys. Rev. Lett., 2009, vol. 102, no. 18, article 183903 (4 pages). 168. Kundtz, N., Roberts, D.A., Allen, J., Cummer, S., and Smith, D.R., Optical Source Transformations, Opt. Express, 2008, vol. 16, no. 26, pp. 2121521222. 169. Cummer, S.A., Kundtz, N., and Popa, B. I., Electro magnetic Surface and Line Sources under Coordinate Transformations, Phys. Rev. A: At., Mol., Opt. Phys., 2009, vol. 80, no. 3, article 033820 (7 pages). 170. Leonhardt, U. and Tyc, T., Superantenna Made of Transformation Media, New J. Phys., 2008, vol. 10, article 115026 (9 pages).

553

171. Greenleaf, A., Kurylev, Y., Lassas, M., and Uhlmann, G., Invisibility and Inverse Problems, Bull. Am. Math. Soc., 2008, vol. 46, no. 1, pp. 5597. 172. Cummer, S.A., Rahm, M., and Schurig, D., Material Parameters and Vector Scaling in Transformation Acoustics, New J. Phys., 2008, vol. 10, article 115025 (12 pages). 173. Li, J. and Chan, C.T., Double Negative Acoustic Metamaterial, Phys. Rev. E: Stat., Nonlinear, Soft Mat ter Phys., 2004, vol. 70, no. 5, article 055602 (R) (4 pages). 174. Lee, S.H., Park, C.M., Seo, Y.M., Wang, Z.G., and Kim, C.K., Composite Acoustic Medium with Simul taneously Negative Density and Modulus, Phys. Rev. Lett., 2010, vol. 104, no. 5, article 054301 (4 pages). 175. Ambati, M., Fang, N., Sun, C., and Zhang, X., Sur face Resonant States and Superlensing in Acoustic Metamaterials, Phys. Rev. B: Condens. Matter, 2007, vol. 75, no. 19, article 195447 (5 pages). 176. Chen, H. and Chan, C.T., Acoustic Cloaking in Three Dimensions Using Acoustic Metamaterials, Appl. Phys. Lett., 2007, vol. 91, no. 18, article 183518 (3 pages). 177. Zhang, S., Yin, L., and Fang, N., Focusing Ultra sound with an Acoustic Metamaterial Network, Phys. Rev. Lett., 2009, vol. 102, no. 19, article 194301 (4 pages). 178. Li, J., Fok, L., Yin, X., Bartal, G., and Zhang, X., Experimental Demonstration of an Acoustic Magnifying Hyperlens, Nat. Mater., 2009, vol. 8, no. 12, pp. 931 934.

GLASS PHYSICS AND CHEMISTRY

Vol. 36

No. 5

2010