## Abstract

Metamaterial absorbers typically consist of a metamaterial layer, a dielectric spacer layer, and a metallic ground plane. We have investigated the dependence of the metamaterial absorption maxima on the spacer layer thickness and the reflection coefficient of the metamaterial layer obtained in the absence of the ground plane layer. Specifically, we employ interference theory to obtain an analytical expression for the spacer thickness needed to maximize the absorption at a given frequency. The efficacy of this simple expression is experimentally verified at terahertz frequencies through detailed measurements of the absorption spectra of a series of metamaterials structures with different spacer thicknesses. Using an array of split-ring resonators (SRRs) as the metamaterial layer and SU8 as the spacer material we observe that the absorption peaks redshift as the spacer thickness is increased, in excellent agreement with our analysis. Our findings can be applied to guide metamaterial absorber designs and understand the absorption peak frequency shift of sensors based on metamaterial absorbers.

© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

## Corrections

22 January 2018: Typographical corrections were made to the abstract, body text, and funding section.

## 1. Introduction

Metamaterials (MMs) are engineered materials with subwavelength unit cells periodically patterned to exhibit extraordinary electromagnetic properties that do not exist in natural materials, such as negative refractive index [1–3]. Metamaterial absorbers can achieve near unity absorption by matching the effective wave impedance to that of free space and are typically designed in a three-layer configuration, consisting of a metamaterial layer, an insulating spacer layer, and a metallic ground plane [4,5]. Different absorption characteristics, such as multiple absorption bands, nonlinear response, polarization insensitivity, high quality factor, and incident angle insensitivity can be engineered from microwave to infrared by careful design of the unit cell [6–13]. Tailored perfect absorption at specific frequencies may enable applications ranging from terahertz (THz) imaging and middle infrared imaging to energy harvesting and radiative cooling [14–19]. Moreover, as the absorption peak frequency and amplitude are closely related to the permittivity of the materials surrounding the metamaterial layer, metamaterial absorbers may be precisely configured for sensing applications [20,21].

Various models have been developed to explain resonant near-unity absorption, including effective medium impedance matching [22,23]. In this approach, metamaterial absorbers are treated as homogenous materials with intrinsic complex impedance carefully designed to match free space to minimize surface reflections. The impedance is retrieved using full-wave electromagnetic simulations to obtain the S-parameters. As such, this serves as a rapid means to design and optimize metamaterials absorbers, albeit with some loss of physical insight typical of numerical approaches. Transmission line theory is another well-established approach [24,25], in which the MMs and the spacer material are modeled as equivalent lump impedances in parallel. Although the theory can describe the absorption, the mathematical complexity usually precludes further derivation to reveal deeper relationships. Coupled mode theory, while elegant, also suffers from a similar problem related to mathematical complexity [26,27]. In a more traditional approach, interference theory can be applied to understand the net reflection that results from multiple reflections between the metamaterial layer and the ground plane [28], with relatively simple mathematics offering an opportunity to explore the deeper relationships into the nature of the resonant absorption.

We employ interference theory to obtain an analytical expression for the spacer thickness needed to maximize the absorption at a given frequency. This expression gives the spacer layer thickness in terms of the reflection coefficient of the metamaterial layer in the absence of a ground plane and the complex refractive index of the spacer layer. We find that in the limit of zero spacer thickness (neglecting coupling between the metamaterial layer and the ground plane), the metamaterial absorber maxima would occur at the resonant frequency of the bare (i.e. no ground plane) metamaterial layer. With increasing spacer layer thickness, the metamaterial absorber resonances redshift. Furthermore, with increasing spacer layer thickness, phase delays result in the appearance of additional absorption resonances. To experimentally verify our analysis, we fabricated a series of metamaterial absorbers with SU8 as the spacer material with different thicknesses, and characterized the electromagnetic response using THz time-domain spectroscopy. The experimental results show excellent agreement with the analytical results. Finally, we note that our analytical approach provides insight into the effect of the spacer layer complex permittivity (ε_{r} + iε_{i}) revealing, for example, the expected result that the frequency shift of the metamaterial absorber is primarily tied to changes of ε_{r}.

## 2. Analyzation of the thickness dependence based on interference theory

We have investigated a metamaterial absorber consisting of a single layer of split ring resonators (SRRs) [geometry shown in inset of Fig. 1(b)], a dielectric spacer layer, and a ground plane. If the near field coupling between the ground plane and the MMs is negligible, the net reflection from interference theory is [28]:

where*r*is the net reflection,

*β*=

*nkd*is the one-way phase delay in the spacer, and

*n, k, d*are the refractive index of the spacer material, the wave number in vacuum, and spacer thickness, respectively. Using conventional nomenclature, the reflection and transmission coefficients are

*r*

_{12},

*t*

_{12}(incident from air side) and

*r*

_{21},

*t*

_{21}(incident from substrate side) of the metamaterial layer. Importantly, these are the coefficients of the metamaterial and substrate in the absence of a ground plane, numerically obtained using CST Microwave Studio. Gold was modeled as a lossy metal with conductivity of 4.56 × 10

^{7}S/m (from CST material library) and SU8 was modeled as a dielectric material with relative permittivity of 2.8 × (1 + 0.065i) [29]. Figure 1 plots the amplitude [Fig. 1(a)] and phase [Fig. 1(b)] spectra of the retrieved S-parameters from simulation, which can transfer to the reflection and transmission coefficients as

*r*

_{12}= S

_{11},

*r*

_{21}= S

_{22},

*t*

_{12}= S

_{21}

*n*

^{-1/2},

*t*

_{21}= S

_{12}

*n*

^{1/2}. There are two resonant modes with frequencies of 0.78 and 2.07 THz, corresponding to the LC and dipole resonances. In the following, we refer to these modes as the bare resonances (i.e. the resonant frequencies in the absence of a ground plane).

We now consider the perfect absorber geometry, which includes a ground plane. The transmission (because of the ground plane) is negligible. As such, the frequency dependent absorption is *A* = 1-abs(*r*)^{2}, where *A* is the absorption amplitude and *r* is from Eq. (1). The absorption spectra for different spacer thicknesses are plotted in Fig. 2(a). The bare LC and dipole resonant frequencies are indicated by the vertical dashed lines. The absorption at these frequencies is nearly zero due to the strong reflection and weak transmission at the metamaterial layer.

Typically, metamaterial absorbers are designed with thin (i.e. subwavelength) spacers. As an example, for a 20 μm spacer thickness [highlighted with the horizontal dashed lines in Fig. 2(a)], two absorption peaks are evident and with peak frequencies that are redshifted from the bare values. As the spacer thickness increases there is a clear redshift of the absorption peak. Importantly, as the spacer thickness further increases, additional absorption bands appear at frequencies that are above the bare resonant frequencies. These also redshift with spacer thickness and ultimately cross the bare resonance frequency. This phenomenon results from the phase delay (in the spacer layer) approaching integer multiples of 2π. In the following, an analytically understanding of this response is presented.

The spacer thickness *d*_{0} to achieve the maximum absorption at a specified frequency can be derived from interference theory and is given as:

*a*and

*b*are the real and imaginary parts of

*r*

_{21},

*n*

_{r}and

*n*

_{i}are the real and imaginary parts of the refractive index of the spacer, and

*m*is an integer starting from zero (the derivation is presented in the appendix). Equation (2) is plotted in Fig. 2(b) for different values of

*m*. The line (for a given value of

*m*) gives the required spacer thickness to obtain an absorption resonance maximum at a given frequency. Clearly, with increasing spacer thickness, multiple modes (having different

*m*values) can contribute to the absorption spectrum.

For dielectrics, *n*_{i} is usually much smaller than *n*_{r}. In this limit, Eq. (2) simplifies to:

*m*= 0) the spacer thickness is directly associated with the phase of

*r*

_{21}[arctan(-

*b*/

*a*)] and the wave number (

*k*). We note that arctan(-

*b*/

*a*) only yields positive values at frequencies lower than the bare resonant frequencies of the metamaterial layer. As such, the absorption maxima must lie below the bare MM resonance frequency which is evident in Fig. 2(b) (again, we emphasize that this is for

*m*= 0). With

*m*>0 absorption peaks appear at different values of

*m*as the spacer thickness increases. This is associated with a 2π phase delay throughout the spacer according to Eq. (3). Moreover, since the second term in Eq. (3) is positive instead of zero (for

*m*= 0), the absorption maxima can occur at higher frequencies than the bare resonance, even with negative arctan(-

*b*/

*a*) value. Also, as indicated by the dashed black arrow (Fig. 2), the frequency redshifts with increasing spacer thickness, arising from both the increase in arctan(-

*b*/

*a*) and the decrease in

*k*. We note that this model does not include coupling between the metamaterial layer and the ground plane. The near field coupling between the two metallic layers substantially increases as the metamaterial layer approaches the ground plane, resulting in a saturated absorption peak frequency lower than the value predicted by the

*m*= 0 curve [26,30,31]. However, it is expected that this only occurs for extremely thin spacer layers as the high-order evanescent waves die off quickly with increase in spacer thickness.

## 3. Experiments and results

To verify the analysis presented above, metamaterial absorbers were fabricated [see unit cell structure and lateral dimensions in the inset of Fig. 1(b)] utilizing SU8 as the spacer material with different thicknesses (10 μm, 16 μm, 28 μm, 70 μm, 90 μm, and 105 μm). E-beam evaporation was employed to deposit Ti/Au/Ti with thickness of 20/200/20 nm as the ground plane on a silicon substrate. The first layer of titanium was applied as an adhesion layer for the subsequent gold deposition, while the second layer of titanium was utilized to promote the adhesion between the SU8 and the ground plane. Next, SU8 with different thicknesses was spin coated on the ground plane. SU8 3010 was applied to fabricate the samples with spacer thicknesses of 10 μm, 16 μm, and 28 μm. SU8 2035 was applied to fabricate the samples with spacer thicknesses of 70 μm, 90 μm, and 105 μm. The desired thickness was achieved through calibrated control of the spinner rotation. The hard bake temperature was set to 95 °C to prevent the peel-off of SU8. AZ5214E was applied to photolithography the pattern of the MMs on the SU8 films, followed by e-beam evaporation to deposit Ti/Au with thicknesses of 20/200 nm and lift-off process to obtain the final structure. A fabricated sample and the microscope image of the unit cells are shown in Fig. 3.

The metamaterial absorbers were characterized with reflection based THz time domain spectroscopy (THz-TDS). THz radiation was focused onto the sample at normal incidence. The polarization direction is indicated by the white arrow in Fig. 3(b). The absorption spectra obtained numerically, theoretically, and experimentally were plotted in Fig. 4(a) showing high degree of agreement. The discrepancy of baseline of the absorption spectra in Fig. 4(a) is due to the imperfect referencing in a reflection geometry.

In Fig. 5(a), the absorption peak frequencies from experiment, interference theory, and simulation are plotted together with the m-curves from Eq. (2). There is clearly excellent agreement between experiment and theory in addition to consistent results between Eq. (2), the full interference approach, and simulations. The simulated spectra of the metamaterial absorbers are presented in Fig. 5(b) showing the dependence with increasing spacer thickness. The bare LC resonance frequency is denoted with a vertical green dashed line [Figs. 5(a) and (b)]. Due to the strong reflection, the magnitude of the absorption is strongly diminished at the bare resonance frequency. With thin spacer thicknesses (10 μm, 16 μm, and 28 μm), the absorption peak frequency is lower than the bare resonant frequency and, as discussed above, redshifts as the thickness increases. When the spacer thickness is large (70 μm, 90 μm, and 105 μm), the absorption peak frequency is higher than the bare resonant frequency approaching the resonant frequency as the spacer thickness increases.

Figure 6 depicts the SRRs current (70 μm spacer thickness) at the resonance frequencies to illustrate the character of the resonant modes. It is evident that the LC (dipole) mode contributes to the absorption band at the frequencies of 0.4376 THz and 0.9578 THz (1.586 THz). We note that the mode character is not determined by the value of *m*, but rather by the bare resonant frequency, which means with increasing frequency the contributing resonant mode for the absorption gradually transfers from LC to dipolar, as indicated by the shaded color in Fig. 2(b).

Using Eq. (2) it is also possible to investigate the effect of the spacer permittivity on the peak absorption frequency. The m-curves (for *m* = 0) were numerically calculated for different permittivity combinations of the spacer material. Figure 7(a) plots the m-curves obtained upon modifying the real part of the permittivity (ε_{r}) of the spacer material (with the imaginary permittivity ε_{i} = 0.182). As ε_{r} increases, the resonant frequencies of both the LC and dipole modes redshift. This trend can be understood by the increase in the equivalent capacitance for larger ε_{r}. Figure 7(b) plots the zoomed in m-curves across a relatively small frequency scale obtained by changing ε_{i} (with ε_{r} = 2.8). Compared with the frequency shift caused by ε_{r}, the frequency shift caused by ε_{i} is miniscule as further supported by the high degree of overlap among the m-curves shown by the inset over a larger frequency range. Thus, the absorption peak frequency shift caused by the variation of imaginary permittivity can be neglected. Nevertheless, a change of ε_{i} will modify the absorption magnitude.

## 4. Conclusion

In summary, we mathematically derived the relationship between the absorption peak frequencies of the metamaterial absorbers with the resonant frequencies of the MMs and the spacer thickness. We proved that the absorption peak frequencies are closely related to the resonant frequencies of the MMs with a frequency shift determined by the phase delay in the spacer material. Moreover, we utilized SU8 as the spacer material, fabricated metamaterial absorbers with different spacer thicknesses, and characterized them with THz TDS. The measurement results exhibited high agreement with the analytical and simulation results. Also, we employed our findings to analyze the contribution of the spacer permittivity change to the absorption peak frequency shift, and revealed that the real part of the permittivity predominates the frequency shift, rather than the imaginary part. The conclusions derived above can be directly applied to any metamaterial absorber configured with metamaterial-spacer-ground plane structure, can be utilized to understand the response of sensors based on such devices, and can guide the design and optimization process.

## Appendix

In the derivation below, *a* and *b* are the real and imaginary parts of *r*_{21}, *β* is the one-way phase delay in the spacer layer with *β = β*_{r} + i*β*_{i}, *d* is the spacer thickness, and *m* is an integer. The net reflection can be expressed with Eq. (1). According to law of cosines, the amplitude of the total reflection can be written as:

*θ*is the angle between the first and secondary reflection, and

*A*and

*B*are the amplitude of the first and secondary reflection, which can be written as:

*d*is the spacer thickness. Around the perfect absorption condition, the equation is valid when (

*B*is only equal to

*A*at the perfect absorption point) d

*B*/d

*d*= 0, which can be expanded as:

*n*

_{i}

*k*exp(2

*β*

_{i}):

*β*

_{r}):

*β*

_{r}equals to

*m*π:

*β*

_{r}is near to

*m*π:

*β*

_{r}=

*n*

_{r}

*kd*

_{0}, the spacer thickness can be expressed as:

*d*

_{0}represents the spacer thickness to obtain an absorption maximum. We note that the expression was derived in the vicinity of perfect absorption, and may give larger error at the frequencies far from the perfect absorption frequency. Nevertheless, the derived result can reasonably describe the frequency shift with the change of the spacer thickness as shown in Fig. 5(a).

We also note that the m-curves in this manuscript only include the case in which the first and secondary reflections cancel out instead of adding up by only considering the round-trip phase delay [phase of *r*_{21} add with delay in the spacer as Eq. (3)]. This is because we only consider the situations where the phase delay gives 2*m*π (cancel out) as opposed to (2*m*+1)π (add up) [Eq. (14)]. To support this fact, we plot the absorption and phase of the absorber at three frequency points with different spacer thicknesses in Fig. 8. The figure clearly shows that at the spacer thickness with peak absorption [estimated with Eq. (2)] the phase difference between the first and secondary reflection approximately π.

## Funding

National Science Foundation (NSF) Grant No. ECCS-1309835; Army Research Office under ARO W911NF-16-1-0361.

## Acknowledgment

The authors would like to thank Boston University Photonics Center for technical support.

## References and links

**1. **D. R. Smith, W. J. Padilla, D. C. Vier, S. C. Nemat-Nasser, and S. Schultz, “Composite medium with simultaneously negative permeability and permittivity,” Phys. Rev. Lett. **84**(18), 4184–4187 (2000). [CrossRef] [PubMed]

**2. **D. R. Smith, J. B. Pendry, and M. C. K. Wiltshire, “Metamaterials and negative refractive index,” Science **305**(5685), 788–792 (2004). [CrossRef] [PubMed]

**3. **X. Zhang, N. Xu, K. Qu, Z. Tian, R. Singh, J. Han, G. S. Agarwal, and W. Zhang, “Electromagnetically induced absorption in a three-resonator metasurface system,” Sci. Rep. **5**(1), 10737 (2015). [CrossRef] [PubMed]

**4. **N. I. Landy, S. Sajuyigbe, J. J. Mock, D. R. Smith, and W. J. Padilla, “Perfect metamaterial absorber,” Phys. Rev. Lett. **100**(20), 207402 (2008). [CrossRef] [PubMed]

**5. **H. Tao, C. M. Bingham, A. C. Strikwerda, D. Pilon, D. Shrekenhamer, N. I. Landy, K. Fan, X. Zhang, W. J. Padilla, and R. D. Averitt, “Highly flexible wide angle of incidence terahertz metamaterial absorber: Design, fabrication, and characterization,” Phys. Rev. B **78**(24), 241103 (2008). [CrossRef]

**6. **X. Shen, Y. Yang, Y. Zang, J. Gu, J. Han, W. Zhang, and T. J. Cui, “Triple-band terahertz metamaterial absorber: Design, experiment, and physical interpretation,” Appl. Phys. Lett. **101**(15), 154102 (2012). [CrossRef]

**7. **H. Tao, C. M. Bingham, D. Pilon, K. Fan, A. C. Strikwerda, D. Shrekenhamer, W. J. Padilla, X. Zhang, and R. D. Averitt, “A dual band terahertz metamaterial absorber,” J. Phys. D Appl. Phys. **43**(22), 225102 (2010). [CrossRef]

**8. **X. Zhao, J. Zhang, K. Fan, G. Duan, G. D. Metcalfe, M. Wraback, X. Zhang, and R. D. Averitt, “Nonlinear terahertz metamaterial perfect absorbers using GaAs,” Photon. Res. **4**(3), A16–A21 (2016). [CrossRef]

**9. **J. Grant, Y. Ma, S. Saha, A. Khalid, and D. R. S. Cumming, “Polarization insensitive, broadband terahertz metamaterial absorber,” Opt. Lett. **36**(17), 3476–3478 (2011). [CrossRef] [PubMed]

**10. **D. T. Viet, N. T. Hien, P. V. Tuong, N. Q. Minh, P. T. Trang, L. N. Le, Y. P. Lee, and V. D. Lam, “Perfect absorber metamaterials: Peak, multi-peak, and broadband absorption,” Opt. Commun. **322**, 209–213 (2014). [CrossRef]

**11. **Y. Ma, Q. Chen, J. Grant, S. C. Saha, A. Khalid, and D. R. S. Cumming, “A terahertz polarization insensitive dual band metamaterial absorber,” Opt. Lett. **36**(6), 945–947 (2011). [CrossRef] [PubMed]

**12. **S. Bhattacharyya and K. V. Srivastava, “Triple band polarization-independent ultra-thin metamaterial absorber using electric field-driven LC resonator,” J. Appl. Phys. **115**(6), 064508 (2014). [CrossRef]

**13. **S. K. Sharma, S. Ghosh, and K. V. Srivastava, “An ultra-thin triple-band polarization-insensitive metamaterial absorber for S, C and X band applications,” Appl. Phys., A Mater. Sci. Process. **122**(12), 1071 (2016). [CrossRef]

**14. **C. M. Watts, D. Shrekenhamer, J. Montoya, G. Lipworth, J. Hunt, T. Sleasman, S. Krishna, D. R. Smith, and W. J. Padilla, “Terahertz compressive imaging with metamaterial spatial light modulators,” Nat. Photonics **8**(8), 605–609 (2014). [CrossRef]

**15. **N. I. Landy, C. M. Bingham, T. Tyler, N. Jokerst, D. R. Smith, and W. J. Padilla, “Design, theory, and measurement of a polarization-insensitive absorber for terahertz imaging,” Phys. Rev. B **79**(12), 125104 (2009). [CrossRef]

**16. **T. Maier and H. Brueckl, “Multispectral microbolometers for the midinfrared,” Opt. Lett. **35**(22), 3766–3768 (2010). [CrossRef] [PubMed]

**17. **X. Liu, T. Tyler, T. Starr, A. F. Starr, N. M. Jokerst, and W. J. Padilla, “Taming the blackbody with infrared metamaterials as selective thermal emitters,” Phys. Rev. Lett. **107**(4), 045901 (2011). [CrossRef] [PubMed]

**18. **M. M. Hossain, B. Jia, and M. Gu, “A metamaterial emitter for highly efficient radiative cooling,” Adv. Opt. Mater. **3**(8), 1047–1051 (2015). [CrossRef]

**19. **A. Vora, J. Gwamuri, N. Pala, A. Kulkarni, J. M. Pearce, and D. O. Güney, “Exchanging Ohmic losses in metamaterial absorbers with useful optical absorption for photovoltaics,” Sci. Rep. **4**(1), 4901 (2015). [CrossRef] [PubMed]

**20. **N. Liu, M. Mesch, T. Weiss, M. Hentschel, and H. Giessen, “Infrared perfect absorber and its application as plasmonic sensor,” Nano Lett. **10**(7), 2342–2348 (2010). [CrossRef] [PubMed]

**21. **X. Hu, G. Xu, L. Wen, H. Wang, Y. Zhao, Y. Zhang, D. R. S. Cumming, and Q. Chen, “Metamaterial absorber integrated microfluidic terahertz sensors,” Laser Photonics Rev. **10**(6), 962–969 (2016). [CrossRef]

**22. **H. Tao, N. I. Landy, C. M. Bingham, X. Zhang, R. D. Averitt, and W. J. Padilla, “A metamaterial absorber for the terahertz regime: Design, fabrication and characterization,” Opt. Express **16**(10), 7181–7188 (2008). [CrossRef] [PubMed]

**23. **C. M. Watts, X. Liu, and W. J. Padilla, “Metamaterial electromagnetic wave absorbers,” Adv. Mater. **24**(23), OP98 (2012). [PubMed]

**24. **F. Costa, S. Genovesi, A. Monorchio, and G. Manara, “A circuit-based model for the interpretation of perfect metamaterial absorbers,” IEEE Trans. Antenn. Propag. **61**(3), 1201–1209 (2013). [CrossRef]

**25. **Q. Y. Wen, Y. S. Xie, H. W. Zhang, Q. H. Yang, Y. X. Li, and Y. L. Liu, “Transmission line model and fields analysis of metamaterial absorber in the terahertz band,” Opt. Express **17**(22), 20256–20265 (2009). [CrossRef] [PubMed]

**26. **S. Ma, S. Xiao, and L. Zhou, “Resonant modes in metal/insulator/metal metamaterials: An analytical study on near-field couplings,” Phys. Rev. B **93**(4), 045305 (2016). [CrossRef]

**27. **C. Qu, S. Ma, J. Hao, M. Qiu, X. Li, S. Xiao, Z. Miao, N. Dai, Q. He, S. Sun, and L. Zhou, “Tailor the functionalities of metasurfaces based on a complete phase diagram,” Phys. Rev. Lett. **115**(23), 235503 (2015). [CrossRef] [PubMed]

**28. **H.-T. Chen, “Interference theory of metamaterial perfect absorbers,” Opt. Express **20**(7), 7165–7172 (2012). [CrossRef] [PubMed]

**29. **N. Ghalichechian and K. Sertel, “Permittivity and loss characterization of SU-8 films for mmW and terahertz applications,” IEEE Antennas Wirel. Propag. Lett. **14**, 723–726 (2015). [CrossRef]

**30. **L. Huang, D. R. Chowdhury, S. Ramani, M. T. Reiten, S. Luo, A. K. Azad, A. J. Taylor, and H.-T. Chen, “Impact of resonator geometry and its coupling with ground plane on ultrathin metamaterial perfect absorbers,” Appl. Phys. Lett. **101**(10), 101102 (2012). [CrossRef]

**31. **S. A. Tretyakov and C. R. Simovski, “Dynamic model of artificial reactive impedance surfaces,” J. Electromagn. Waves Appl. **17**(1), 131–145 (2003). [CrossRef]