Perhaps all things around us are some types of waves, such as the light waves, acoustic waves, matter waves and even gravitational waves. The absorption or dissipation of the energy carried by these waves is of great importance in many applications ranging from energy harvest and information exchange to radar wave and sound noise control^{18}. Among the various types of waves, the electromagnetic and acoustic waves share many similarities. In the last decades, the two disciplines have borrowed concepts from each other to grow up alternately. For example, the timereversed subdiffraction focusing technique was first developed by the acousticians and then utilized in the optical domain to break the diffraction limit^{912}. On the other hand, the hyperlens was originally created in the optical domain and then introduced to acoustic research^{13, 14}. Recently, the concept of metasurface, an artificially structured thin film with property on demand, has become a hot spot in both the electromagnetic and acoustic regimes^{1518}. It is, however, surprising that many metasurface concepts such as Salisbury and Jaumann absorbers in the electromagnetic design have not been extended into the acoustic domain. Furthermore, although ultrathin and ultrabroadband electromagnetic absorbers have been realized^{19, 20}, similar thin absorbers are still difficult to obtain for acousticians^{4, 17}.
In this paper, we theoretically compared the boundary conditions for electromagnetic and acoustic metasurfaces. Similar to electromagnetic metasurface absorber, we show that various acoustic absorbers can be achieved using the effective impedance of deepsubwavelength perforated rigid plate (may be approximated by steel). In particular, an ultrabroadband and thin acoustic perfect absorber is demonstrated under coherent condition, showing excellent performance comparable to its electromagnetic counterpart.
Results Analogy between the boundary conditions of electromagnetic and acoustic wavesFirst of all, the boundary conditions for electromagnetic metasurface are considered. As shown in Fig. 1(a), the tangential components of the electric and magnetic fields are continuous at the interface between two halfinfinite materials. When a nonmagnetic impedance sheet with zero thickness is placed at this interface^{21}, the boundary conditions should be rewritten as
$ {{\mathit{\boldsymbol{E}}}_{{\text{i}},\parallel }} + {{\mathit{\boldsymbol{E}}}_{{\text{r}},\parallel }} = {{\mathit{\boldsymbol{E}}}_{{\text{t}},\parallel }},\\ {{\mathit{\boldsymbol{H}}}_{{\text{i}}, \parallel }} + {{\mathit{\boldsymbol{H}}}_{{\text{r}}, \parallel }} = {{\mathit{\boldsymbol{H}}}_{{\text{t}}, \parallel }} + {\mathit{\boldsymbol{\hat n}}} \times {Y_{{\text{s, em}}}}{{\mathit{\boldsymbol{E}}}_{{\text{t}}, \parallel }}, $  (1) 
where i, r and t denote the incidence, reflection and transmission,
Since
$\nabla \times {\mathit{\boldsymbol{E}}} =  \mu \frac{{\partial {\mathit{\boldsymbol{H}}}}}{{\partial t}}, $  (2) 
equation (1) can be reduced to be
$ 1 + {r_{{\text{em}}}} = {t_{{\text{em}}}}, \\ {Y_{1, {\text{em}}}}(1  {r_{{\text{em}}}}) = {Y_{2, {\text{em}}}}{t_{{\text{em}}}} + {Y_{s, {\text{em}}}}{t_{{\text{em}}}}, $  (3) 
where r_{em} and t_{em} are the reflection and transmission coefficients defined by the electric fields, Y_{1}_{,}_{em} and Y_{2}_{,}_{em} are the inductances of two materials in the left and right sides.
Similar to the electromagnetic case, the boundary conditions for an acoustic impedance sheet (Fig. 1(b)) has the form of
$ {v_{{\text{i}}, \bot }} + {v_{{\text{r}}, \bot }} = {v_{{\text{t}}, \bot }}, $  (4) 
and
$ {p_{\text{i}}} + {p_{\text{r}}} = {p_{\text{t}}} + {Z_{s{\text{, ac}}}}{v_{{\text{t}}, \bot }}. $  (5) 
where v and p are the velocity and pressure of the sound perturbation. While equation (4) is a direct result of the mass conservation, equation (5) is obtained by defining the effective sheet impedance
$ {Z_{s{\text{, ac}}}} = \frac{{{p_{\text{i}}} + {p_{\text{r}}}  {p_{\text{t}}}}}{{{v_{{\text{t}}, \bot }}}} = \frac{{\Delta p}}{{{{\bar v}_ \bot }}}. $  (6) 
In general, v and p are related through the characteristic impedance
$ {p_{\text{i}}} = {Z_{{\text{1, ac}}}}{v_{{\text{i}}, \bot }}\;, \;\;{p_{\text{r}}} = {Z_{{\text{1, ac}}}}{v_{{\text{r}}, \bot }}\;, \;\;{p_{\text{t}}} = {Z_{{\text{2, ac}}}}{v_{{\text{t}}, \bot }}. $  (7) 
Assuming that the background is filled with air, the impedance can be written as
$ 1 + {r_{{\text{ac}}}} = {t_{{\text{ac}}}}, \\ {Z_{1, {\text{ac}}}}(1  {r_{{\text{ac}}}}) = {Z_{2, {\text{ac}}}}{t_{{\text{ac}}}} + {Z_{s, {\text{ac}}}}{t_{{\text{ac}}}}. $  (8) 
Interestingly, equations (3) and (8) have very similar form, except that the admittance in equation (3) is replaced by the impedance in equation (8). This similarity implies that both the acoustic and electromagnetic problems in metasurfaces can be solved by transfer matrix, along with proper boundary conditions.
Impedance sheet realized by microperforated plateIn the above discussion, an arbitrary sheet impedance is assumed without considering its practical realization. In the following, the effective impedance of microperforated plate (MPP) is utilized in the design of various acoustic absorbers. It should be noted that although the MPP absorbers have been intensively studied after its invention by Maa in 1970s^{2224}, its intrinsic similarity with electromagnetic absorber such as Salisbury and Jaumann absorbers has seldom been studied^{2527}. In this paper, we show that the concept borrowed from electromagnetics is very efficient in the design of acoustic devices.
In general, the effective impedance of a microtube perforated in a rigid plate should be calculated using computational fluid dynamics (CFD)^{23}. However, when the tube is short compared to the wavelength, the effective impedance can be approximately analyzed using the equation of aerial motion
$ i\omega {\rho _0}v  \frac{\eta }{{{r_1}}}\frac{\partial }{{\partial {r_1}}}({r_1}\frac{\partial }{{\partial {r_1}}}v) = \frac{{\Delta p}}{t}, $  (9) 
where v is particle velocity,
$ {Z_{{\text{tube}}}} = \frac{{\Delta p}}{{\bar v}} = i\omega {\rho _0}t{\left[ {1  \frac{2}{{k\sqrt {  i} }}\frac{{{{\text{J}}_1}(\kappa \sqrt {  i} )}}{{{{\text{J}}_0}(\kappa \sqrt {  i} )}}} \right]^{  1}}, $  (10) 
where
$ {Z_{s, {\text{ac}}}} = \frac{{{Z_{{\text{tube}}}}}}{\sigma }, $  (11) 
where
$ {Z_s} = \frac{{t + \beta d}}{t}{Z_{s, {\text{ac}}}}, $  (12) 
where
The normalized impedance at f=10 kHz for different hole diameters is plotted in Fig. 2. Clearly, as the increase of the hole diameter, the impedance drops down rapidly. When the hole is extremely small, the impedance tends to be infinite, with all sound wave being reflected. The frequencydependent impedance of a MPP metasurface is shown in Fig. 3, where the geometric parameters are fixed as d=38.5 μm and b=t=100 μm. Evidently, the resistance is almost independent of the frequency, while the reactance is proportional to the frequency.
From above results, it is clear that a thin microperforated plate bears similar properties with the resistive sheet (thin metal film) in the electromagnetic applications^{25}. Combining with the similarity in boundary conditions, it can be concluded that the acoustic waves can be controlled using electromagneticinspired approaches. Furthermore, it is expected that more complex sheet impedance is possible by combing the microperforated plate with cavities or other resonant structures^{28}.
Acoustic Salisbury and Jaumann absorbersAs the first step to demonstrate the intrinsic connection between the electromagnetic and acoustic metasurface, the acoustictype Salisbury^{25, 29} and Jaumann^{30} absorbers are investigated. As depicted in Fig. 4, the absorption spectra of two optimized absorbers are calculated using transfer matrix method (see methods). Theoretically, the physical mechanism behind the absorption is the multiple interference between successive reflections^{31}. Although only theoretical results are given here, previous experiments agree well with our theory^{24}, indicating that the following discussion is sound and robust.
For the Salisburytype absorber, there is only one microperforated plate (labelled as MPP1) separated from the rigid body with a distance of h_{1} =48 mm. The geometric parameters of MPP1 is d_{1}=38.5 μm and b=t =100 μm. For the Jaumanntype absorber, there are four MPPs placed successively with distance h_{1}=h_{2}=h_{3}=h_{4}=12 mm. The thickness t and period b are equal for all these MPPs, while the hole diameters are d_{1}=38.5 μm, d_{2}=40 μm, d_{3}=45 μm and d_{4}=50 μm. Note that the onelayer Salisbury and fourlayer Jaumann structures are two representing examples. For the two and threelayer structures, the bandwidth would be larger than the onelayer but smaller than the fourlayer cases. Similar to the electromagnetic case, the thickness mainly determines the absorption frequency. As the thickness becomes larger, the frequency will become smaller.
As illustrated in Fig. 5, the dependence of the acoustic Jaumann absorber on the incidence angle θ is also investigated. Similar to its electromagnetic counterpart, the absorption efficiency would decrease as the rise of the incidence angle although most is larger than 90% for θ < 60°. In addition, we noted that the angle dependence is similar to the transverse electric (TE) mode, rather than the transverse magnetic (TM) mode, as can be derived from the boundary conditions.
Anyway, the results shown in Figs. 4 and 5 resemble that for classic electromagnetic absorbers^{1}. Consequently, we expect that the theory of the recent developed electromagnetic absorbers^{3, 32} can be leveraged to design acoustic absorbers in a more sophisticated way.
Coherent perfect absorption of acoustic waveDifferent from traditional absorbers, coherent perfect absorption (CPA) is a process that requires two or more coherent inputs^{19, 20, 3337}. Owing to the revision of the operation condition, previous limitation on the bandwidth and absorber thickness, as set by the casualty and KramersKronig relations^{38}, can be eliminated^{19, 39}. This broadband coherent absorption characteristic seems a surprise since CPA was thought as a timereversed process of lasing, which is intrinsically narrowband^{40}. Nevertheless, from the view point of electrodynamics, this broadband absorption can be indeed treated as the timereversed process of the radiating current sheet^{19}, which is not limited in bandwidth. Although CPA is of particular importance for electromagnetic applications, its realization in acoustic region is hampered by the absence of proper lossy materials^{17}. Although there are some attempts to design acoustic CPA^{41}, the results seems not so satisfactory because only narrowband absorption can be achieved.
Figure 6 illustrated the absorption spectrum under CPA condition. The hole diameter of the MPP is 32.1 μm, while the thickness and period are both 100 μm. As we expected, the absorption can be as high as 99% at frequencies below 10 kHz, which is very similar to the ultrabroadband electromagnetic CPA^{19}. It should be noted that the normalized impedance for the MPP at CPA condition is nearly twice that of air, while in the electromagnetic case the impedance is only half. This difference is a direct result of the different definition of impedance in acoustics and electromagnetics. Furthermore, since the amplitude of the acoustic wave is defined using the pressure p, the CPA condition requires s=1 (asymmetric mode), which is also contrary to the electromagnetic CPA^{19}.
One of the intrinsic advantages of CPA over traditional absorbers lies in the fact that the absorption can be easily tuned by varying the phase difference between the left and right inputs (l). Similar to the electromagnetic counterpart^{19}, the absorption may be written as:
$ A = 1  {\sin ^2}(\frac{{\omega l}}{{2{v_0}}}), $  (13) 
where v_{0} = 343 m/s is the velocity of sound in air.
As shown in Fig. 7(a), when the path difference is fixed to be 0.2 m, the absorption spectra vary rapidly with the sound frequency. Note that the absorption at very low frequency is always near 100%, implying that this technique is very efficient for the absorption at low frequency. Figure 7(b) depicts the absorption at 10 kHz for various path difference, where the ratio of the minimum to maximum can be as smaller as 2e5, which is beneficial to the coherent acoustic modulators. Similar to the concept of "controlling light with light"^{42}, the acoustic CPA can be also leveraged to realize "controlling sound with sound" on metasurfaces.
DiscussionsIn summary, we have unified the theory of electromagnetic and acoustic absorbers in the context of metasurfaces. It is demonstrated unambiguously that the theory of electromagnetic absorbers can provide useful guidance for acoustic researches. The ultrabroadband acoustic CPA proves that previous judgment about sound absorption, especially at low frequency, is not accurate^{17}.
It should be pointed out that the impedance of the MPP is mainly determined by the radius of the deepsubwavelengthscale holes, which are also of particular importance in electromagnetic scenarios such as the extraordinary optical transmission^{43} and frequency selective surfaces^{44}. As the operational frequency rises, the holes need to be scaled down to maintain the required impedance and there will be huge challenges to fabricate such small holes in a large area. Meanwhile, the accuracy of Maa's model needs further experimental investigation at higher frequency. Nevertheless, there is no problem if we focus on the lowfrequency regime, which is just the difficult point of traditional approaches. Potential applications may be found in the noise control in both air and water environments.
Method of transfer matrix for the multilayered acoustic absorberSimilar to the electromagnetic case, the multilayered acoustic structure can be calculated efficiently using the transfer matrix. As depicted in Fig. 8, the coefficients of the counterpropagating acoustic waves are denoted as A and B, while the two pairs of quantities can be written as
$ \left[ {\begin{array}{*{20}{c}} {{A_{m + 1}}} \\ {{B_{m + 1}}} \end{array}} \right] = \frac{1}{{2{Z_{m + 1}}}}\left[ {\begin{array}{*{20}{c}} {{Z_{m + 1}} + {Z_m} + {Z_{s, m}}}&{{Z_{m + 1}}  {Z_m} + {Z_{s, m}}} \\ {{Z_{m + 1}}  {Z_m}  {Z_{s, m}}}&{{Z_0} + {Z_1}  {Z_{s, m}}} \end{array}} \right]\\ \;\;\;\;\;\;\;\;\;\;\;\; \cdot \left[ {\begin{array}{*{20}{c}} {\exp (  {\text{i}}k{h_m})}&0 \\ 0&{\exp ({\text{i}}k{h_m})} \end{array}} \right]\left[ {\begin{array}{*{20}{c}} {{A_m}} \\ {{B_m}} \end{array}} \right], $  (14) 
where
$ {{\mathit{\boldsymbol{M}}}_{1, m}} = \frac{1}{{2{Z_{m + 1}}}}\left[ {\begin{array}{*{20}{c}} {{Z_{m + 1}} + {Z_m} + {Z_{s, m}}}&{{Z_{m + 1}}  {Z_m} + {Z_{s, m}}} \\ {{Z_{m + 1}}  {Z_m}  {Z_{s, m}}}&{{Z_0} + {Z_1}  {Z_{s, m}}} \end{array}} \right] $  (15) 
is the transition matrix of the metasurface, while
$ {{\mathit{\boldsymbol{M}}}_{2, m}} = \left[ {\begin{array}{*{20}{c}} {\exp (  {\text{i}}k{h_m})}&0 \\ 0&{\exp ({\text{i}}k{h_m})} \end{array}} \right] $  (16) 
is the propagation matrix of the m^{th} layer. Here
Obviously, the amplitudes of acoustic waves at the outmost layer can be calculated from the innermost directly. Since there is a rigid body acting as a perfect reflector, the final form of the transfer matrix can be written as
$ \left[ {\begin{array}{*{20}{c}} {{A_{N + 1}}} \\ {{B_{N + 1}}} \end{array}} \right] = {{\mathit{\boldsymbol{M}}}_N}{{\mathit{\boldsymbol{M}}}_{N  1}} \cdot \cdot \cdot \cdot \cdot \cdot {{\mathit{\boldsymbol{M}}}_2}{{\mathit{\boldsymbol{M}}}_1}\left[ {\begin{array}{*{20}{c}} 1 \\ {  1} \end{array}} \right], $  (17) 
where M_{m}=M_{1,m}M_{2,m} is the transfer matrix of a pair of air and metasurface. The reflection and absorption coefficients can be then calculated as
$ Reflection = \frac{{{B_{N + 1}}}}{{{A_{N + 1}}}}, $  (18) 
$ {\rm{and}}~~~ Absorption = 1  {\left( {\frac{{{B_{N + 1}}}}{{{A_{N + 1}}}}} \right)^2}. $  (19) 
We acknowledge the financial support by 973 Program of China under contract No. 2013CBA01700 and the National Natural Science Foundation of China under contract No. 61622509 and 61575203.
Competing interestsThe authors declare no competing financial interests.
1. 
Knott E F, Shaeffer J F, Tuley M T. Radar Cross Section 2nd ed (SciTech Publishing, Raleigh, North Carolina, 2004).

2. 
Hao J M, Wang J, Liu X L, Padilla W J, Zhou L et al. High performance optical absorber based on a plasmonic metamaterial. Appl Phys Lett
96, 251104 (2010) [Crossref] 
3. 
Feng Q, Pu M B, Hu C G, Luo X G. Engineering the dispersion of metamaterial surface for broadband infrared absorption. Opt Lett
37, 21332135 (2012) [Crossref] 
4. 
Mei J, Ma G C, Yang M, Yang Z Y, Wen W J et al. Dark acoustic metamaterials as super absorbers for lowfrequency sound. Nat Commun
3, 756 (2012) [Crossref] 
5. 
Vora A, Gwamuri J, Pala N, Kulkarni A, Pearce J M et al. Exchanging ohmic losses in metamaterial absorbers with useful optical absorption for photovoltaics. Sci Rep
4, 4901 (2014) [Crossref] 
6. 
Song M W, Yu H L, Hu C G, Pu M B, Zhang Z J et al. Conversion of broadband energy to narrowband emission through doublesided metamaterials. Opt Express
21, 3220732216 (2013) [Crossref] 
7. 
Cui Y X, He Y R, Jin Y, Ding F, Yang L et al. Plasmonic and metamaterial structures as electromagnetic absorbers. Laser Photonics Rev
8, 495520 (2014) [Crossref] 
8. 
Luo X G. Principles of electromagnetic waves in metasurfaces. Sci China Phys Mech Astron
58, 594201 (2015) [Crossref] 
9. 
de Rosny J, Fink M. Overcoming the diffraction limit in wave physics using a timereversal mirror and a novel acoustic sink. Phys Rev Lett
89, 124301 (2002) [Crossref] 
10. 
Lerosey G, de Rosny J, Tourin A, Fink M. Focusing beyond the diffraction limit with farfield time reversal. Science
315, 11201122 (2007) [Crossref] 
11. 
Chen L W, Zhou Y, Wu M X, Hong M H. Remotemode microsphere nanoimaging: new boundaries for optical microscopes. OptoElectron Adv
1, 170001 (2018) 
12. 
Qin F, Hong M H. Breaking the diffraction limit in far field by planar metalens. Sci China Phys Mech Astron
60, 044231 (2017) [Crossref] 
13. 
Jacob Z, Alekseyev L V, Narimanov E. Optical hyperlens: Farfield imaging beyond the diffraction limit. Opt Express 14, 82478256 (2006).

14. 
Li J, Fok L, Yin X B, Bartal G, Zhang X. Experimental demonstration of an acoustic magnifying hyperlens. Nat Mater
8, 931934 (2009) [Crossref] 
15. 
Kildishev A V, Boltasseva A, Shalaev V M. Planar photonics with metasurfaces. Science
339, 1232009 (2013) [Crossref] 
16. 
Yu N F, Capasso F. Flat optics with designer metasurfaces. Nat Mater
13, 139150 (2014) [Crossref] 
17. 
Ma G C, Yang M, Xiao S W, Yang Z Y, Sheng P. Acoustic metasurface with hybrid resonances. Nat Mater
13, 873878 (2014) [Crossref] 
18. 
Luo X G. Subwavelength optical engineering with metasurface waves. Adv Opt Mater
6, 1701201 (2018) [Crossref] 
19. 
Pu M B, Feng Q, Wang M, Hu C G, Huang C et al. Ultrathin broadband nearly perfect absorber with symmetrical coherent illumination. Opt Express
20, 22462254 (2012) [Crossref] 
20. 
Li S C, Luo J, Anwar S, Li S, Lu W X et al. Broadband perfect absorption of ultrathin conductive films with coherent illumination: Superabsorption of microwave radiation. Phys Rev B
91, 220301 (2015) [Crossref] 
21. 
Pu M B, Hu C G, Huang C, Wang C T, Zhao Z Y et al. Investigation of Fano resonance in planar metamaterial with perturbed periodicity. Opt Express
21, 9921001 (2013) [Crossref] 
22. 
Maa DY. Potential of microperforated panel absorber. J Acoust Soc Am
104, 28612866 (1998) [Crossref] 
23. 
Herdtle T, Bolton J S, Kim N N, Alexander J H, Gerdes R W. Transfer impedance of microperforated materials with tapered holes. J Acoust Soc Am
134, 4752 (2013) [Crossref] 
24. 
Qian Y J, Kong D Y, Liu S M, Sun S M, Zhao Z. Investigation on microperforated panel absorber with ultramicro perforations. Appl Acoust
74, 931935 (2013) [Crossref] 
25. 
Chambers B. Optimum design of a salisbury screen radar absorber. Electron Lett
30, 13531354 (1994) [Crossref] 
26. 
Knott E F, Langseth K. Performance degradation of Jaumann absorbers due to curvature. IEEE Trans Antennas Propag
28, 137139 (1980) [Crossref] 
27. 
Duan Y T, Luo J, Wang G H, Hang Z H, Hou B et al. Theoretical requirements for broadband perfect absorption of acoustic waves by ultrathin elastic metafilms. Sci Rep
5, 12139 (2015) [Crossref] 
28. 
Cheng Y, Zhou C, Yuan B G, Wu D J, Wei Q et al. Ultrasparse metasurface for high reflection of lowfrequency sound based on artificial Mie resonances. Nat Mater
14, 10131019 (2015) [Crossref] 
29. 
Smith F C. Design principles of broadband adaptive Salisbury screen absorber. Electron Lett
38, 10521054 (2002) [Crossref] 
30. 
Munk B A, Munk P, Pryor J. On designing Jaumann and circuit analog absorbers (CA absorbers) for oblique angle of incidence. IEEE Trans Antennas Propag
55, 186193 (2007) [Crossref] 
31. 
Pu M B, Chen P, Wang Y Q, Zhao Z Y, Wang C T et al. Strong enhancement of light absorption and highly directive thermal emission in graphene. Opt Express
21, 1161811627 (2013) [Crossref] 
32. 
Akselrod G M, Huang J N, Hoang T B, Bowen P T, Su L et al. Largearea metasurface perfect absorbers from visible to nearinfrared. Adv Mater
27, 80288034 (2015) [Crossref] 
33. 
Chong Y D, Ge L, Cao H, Stone A D. Coherent perfect absorbers: timereversed lasers. Phys Rev Lett
105, 053901 (2010) [Crossref] 
34. 
Pu M B, Feng Q, Hu C G, Luo X G. Perfect absorption of light by coherently induced plasmon hybridization in ultrathin metamaterial film. Plasmonics
7, 733738 (2012) [Crossref] 
35. 
Li S C, Duan Q, Li S, Yin Q, Lu W X et al. Perfect electromagnetic absorption at oneatomthick scale. Appl Phys Lett
107, 181112 (2015) [Crossref] 
36. 
Papaioannou M, Plum E, Valente J, Rogers E T F, Zheludev N I. Twodimensional control of light with light on metasurfaces. Light Sci Appl
5, e16070 (2016) [Crossref] 
37. 
Li X, Pu M B, Wang Y Q, Ma X L, Li Y et al. Dynamic control of the extraordinary optical scattering in semicontinuous 2D metamaterials. Adv Opt Mater
4, 659663 (2016) [Crossref] 
38. 
Rozanov K N. Ultimate thickness to bandwidth ratio of radar absorbers. IEEE Trans Antennas Propag
48, 12301234 (2000) [Crossref] 
39. 
Wang D C, Zhang L C, Gu Y H, Mehmood M Q, Gong Y D et al. Switchable ultrathin quarterwave plate in terahertz using active phasechange metasurface. Sci Rep
5, 15020 (2015) [Crossref] 
40. 
Wan W J, Chong Y D, Ge L, Noh H, Stone A D et al. Timereversed lasing and interferometric control of absorption. Science
331, 889892 (2011) [Crossref] 
41. 
Wei P J, Croenne C, Chu S T, Li J. Symmetrical and antisymmetrical coherent perfect absorption for acoustic waves. Appl Phys Lett
104, 121902 (2014) [Crossref] 
42. 
Zhang J F, MacDonald K F, Zheludev N I. Controlling lightwithlight without nonlinearity. Light Sci Appl
1, e18 (2012) [Crossref] 
43. 
Ebbesen T W, Lezec H J, Ghaemi H F, Thio T, Wolff P A. Extraordinary optical transmission through subwavelength hole arrays. Nature
391, 667669 (1998) [Crossref] 
44. 
Munk B A. Frequency Selective Surfaces: Theory and Design (Wiley, New York, 2000).
