An ultra-broadband perfect absorber based on graded-index mechanism is designed and fabricated. The perfect absorber is comprised of a heavily-doped silicon absorption substrate and a flat six-layer antireflective structure. The refractive index of each layer was widely tuned by hollow polystyrene microsphere and TiO2 nanoparticle dopants, which can offer a gradually changed refractive index profile from 1.3 to 2.9. The experimental results show that 98% absorption can be achieved within the range of 0.1–20 THz. Moreover, the high absorption efficiency as well as the ultra-broad range can maintain for incident angle from 0 to 75° by the theoretical simulation.
© 2016 Optical Society of America
Achieving perfect broadband absorption in the terahertz regime has recently garnered attention due to its unique and promising applications , such as thermal detectors, microbolometers, and imaging systems [2–4]. To realize broadband terahertz absorbers, many approaches have been proposed. Using metamaterials, broadband absorption can be realized by overlapping the resonant peaks of different electric ring resonators (ERRs) [5–7]. With additional ERRs, dual band [8,9], triple band [10,11], and multi-band absorbers [12,13] can be realized. However, the bandwidth is limited by the complexity of the absorber and fabrication difficulties. Moreover, broadband terahertz absorption can be achieved using two-dimensional grating structures. By elaborately designing the grating, bandwidths of 1–2 THz  and 0.92–2.4 THz have been realized .
Generally, perfect absorption includes two processes: antireflection (AR) and the absorption of incident electromagnetic waves. The bandwidth property of an absorber is principally decided by the AR bandwidth; thus, increasing AR bandwidth is a vital issue. In addition to the metamaterial, the grating, λ/4 coating (where λ is the wavelength in the media) , and a graded-index structure are good candidates for increasing AR. A gradually varying index layer can eliminate Fresnel reflection of the incident electromagnetic wave due to small index differences. This graded-index profile is usually originated using a surface relief structure array. For example, Y.W. Chen et al. proposed a micropyramidal structure with an enhanced bandwidth of 0.2–3.15 THz , and Kim et al. achieved a bandwidth of 0.1–1.9 THz by self-assembly of a hemispherical surface structure . We also attained a bandwidth of 0.2–1.6 THz using a TiO2-polymer composite holes array . However, these graded-index profiles rely on the ratio between air and the material of the surface relief structure. For wavelengths much larger or smaller than the structure, the AR effect will be suppressed.
In this study, we propose a planar ultra-broadband terahertz perfect absorber based on a graded-index mechanism. The device is composed of six flat-layered structures for the antireflection and a heavily doped silicon substrate for the incoming terahertz absorption. To mimic the graded-index profile, the refractive indices of the six layers are varied from1.3 to 2.9 by individually tuning the dopants in the polymer matrix. Unlike the other AR structures mentioned above, there are no relief structures in the plane; therefore, the AR effect can be wavelength independent over a very broad range. The absorption as well as the AR properties is investigated using a terahertz time-domain spectroscopy (THz-TDS) system and Fourier transform infrared (FT-IR) spectroscopy. The incident angle dependency of the absorber is simulated using CST Microwave Studio software. The scattering property was studied utilizing a terahertz quantum cascade laser (QCL) at 4.2 THz. The experimental results show that the proposed planar ultra-broadband terahertz perfect absorber has nearly perfect absorption for incident terahertz waves of at least 0.1–20 THz and is insensitive to the incident angle.
2. Experimental design and device preparation
The schematic diagram of the ultra-broadband terahertz absorber is shown in Fig. 1. The device can be divided into two parts. The top part is the broadband AR layer, which is composed of six flat layers with refractive indices that increase gradually to mimic a graded-index profile; the bottom part is a heavily doped silicon wafer with the thickness 500μm, which is utilized to absorb the incident terahertz wave. The refractive index of each layer is determined using Eq. (1), where n1–6 are the refractive indices of each layer, and nair and nSi are the refractive indices of air and the silicon substrate, respectively. The refractive index profile is shown in Table 1. The thickness of each layer is determined by destructive interference theory in which the frequency was set to 5 THz; the theoretical thicknesses are shown in Table 1. To cover the low terahertz band, a comparably thick layer is identified by 19λ/4n, where λ is 60 μm (5 THz), and n is the refractive index of each layer:
There are three mechanisms that contribute to its ultra-broadband AR effect. For low frequencies, the thickness is smaller than the incident wavelength; in this case, the multilayers act like a graded-index media, and the reflective interface is non-susceptible, which results in low reflection, as shown by the red curve in Fig. 1(a). For high frequencies, the thickness is larger than the wavelength; in this case, the terahertz wave follows the Fresnel equation and is reflected at each layer. However, the refractive index difference for each adjacent layer is so low that most of the terahertz wave can transmit the layers without being reflected, as shown by the blue curve in Fig. 1(c). For some specific frequencies, such as 5 THz, the thickness of each composite layer is an odd multiple of the quarter wavelength of the incident terahertz wave, which is satisfied by destructive interference anti-reflection conditions, as shown by the green curve in Fig. 1(b).
To achieve this multilayer AR structure, we prepared refractive index tunable polymer composites by varying the dopants and doping concentrations. An epoxy (3,4-epoxycyclohexylmethyl 3,4-epoxycyclohexane carboxylate) with a refractive index approximately 1.5 in the terahertz range is used as the host medium for the composites. For the dopant, we used titanium dioxide nanoparticles with a rutile crystal structure (diameter of 20 nm) because it has a relatively large refractive index at terahertz frequencies . According to Rayleigh scattering theory, the scattering loss from the TiO2 particles can be neglected. We prepared the TiO2-epoxy composites as follows. First, the TiO2 nanoparticles are bead-milled with a surface modifier for dispersion in an ethanol solvent to prevent the aggregation of particles . Then, an equivalent proportion of epoxy and hardener (m-xylylenediamine) are dissolved and stirred for 2 hour in the TiO2 suspension. Finally, the composite material is cast on a substrate and cured at 100°C. By controlling the ratio of epoxy and titania, the refractive index of the composites can be controlled. The dependence of refractive index on the weight ratio of TiO2 is shown in Fig. 2.
To characterize the refractive index in the 0.1–1.5 THz range, a THz-TDS system is used. The index of the composite material increases as the doping concentration of the nanoparticles increases. Indices of refraction with values of 1.96, 2.09, 2.41, 2.84, and 3.19 are obtained with corresponding TiO2 weight ratios of 27%, 33%, 44%, 56%, and 67%, respectively, as shown by diamonds in Fig. 2. The refractive index of the TiO2-epoxy composite can be predicted by the complex refractive index (CRI) theory, which is an empirical formula used for dielectric systems with relatively small volume. The explicit description of CRI theory is given asFig. 2. From comparison between the experimental and theoretical results, the CRI model is in good quantitative agreement with experimentally measured refractive indices. According to the fitting curve, composites with indices of 1.8, 2.1, 2.5, and 2.9 (Table 1) were prepared with the weight ratio of 15%, 32%, 46%, and 60%, respectively. The refractive index values beyond the range of 0.1–1.5 THz could differ due to dispersion or absorption, but their changing trend must be the same as the trend in this range.
According to the device design, an extremely low refractive index of 1.2–1.3 is required for the first layer. To achieve such a low refractive index, we used hollow polystyrene (PS) microspheres (Expancel DE@AkazoNobel) as dopants to lower the refractive index of the epoxy resin. The hollow microspheres have diameters from 20 μm to 40 μm, and the thickness of the sphere shell is approximately 0.1 μm; thus, for the wavelength larger or comparable to the diameters, these spheres can greatly decrease the effective refractive index of the composite. The composite preparation procedure is similar to that of the TiO2-epoxy composites. The PS microspheres were first wetted with ethanol, mixed with epoxy and hardener in a weight ratio of 1:1, and stirred 2 h for homogenization. After curing at 60°C, this PS microsphere-epoxy layer has a refractive index of 1.3.
The prepared epoxy composites were successively cast and cured on a heavily doped silicon substrate. The actual thicknesses were measured by a high precision screw micrometer, and listed in Table 1. The cross-section of the fabricated multilayer structure is observed by scanning electron microscopy (SEM), and the result is shown in Fig. 1(d). The refractive index of this material gradually increased from the top layer to the bottom layer. The hollow PS microspheres can be easily identified; however, the other layers are not as clearly distinguished due to their similar states.
3. Results and discussions
The transmittance and reflectance of the fabricated absorber were measured by the THz-TDS system, and the results are shown in Fig. 3. The measuring range was 0.1–1.5 THz. The reflectance was measured by a nearly normal-incident setup; thus, the absorption (A) can be roughly calculated by A = 1-R-T, where the R is the reflectance and T is the transmittance. For a typical frequency of 1.0 THz, the reflectance is 0.13%, the transmittance is 0.36%, and the absorption is 99.51%. From Fig. 3(c), we can see that the absorption of our absorber exceeds 98% in the range of 0.2–1.5 THz. This result demonstrates the excellent absorption performance of our multilayer absorber. The near-zero reflectance illustrates the function of the graded-index composite layer, and the near-zero transmittance demonstrates the blocking effect of the heavily doped silicon substrate. The absorption properties at low frequencies (around 0.1 THz) are not as good as those at high frequencies. For example, at 0.1 THz, the reflectance, transmittance, and absorption are 3.78%, 2.98%, and 93.24%, respectively. At this wavelength range (3 mm), the thickness of the AR layer is far smaller than the incident wavelength; thus, the AR effect decreases quickly as the frequency decreases.
For a broader study of the terahertz properties, the device was also investigated by a FT-IR (Bruker v80) spectroscope equipped with a bolometer detector that offered a valid range from 1.5 to 20 THz. The measured results are presented in Fig. 4. The reflective property was obtained using a 45° angle reflection accessory. From Fig. 4(c), the absorption is more than 99% for the entire measurement range. For a typical frequency at 11 THz, the reflectance, transmittance, and absorption is 0.596%, 0.007%, and 99.397%, respectively.
For practical applications, the incident angle may not be normal to the terahertz absorber; thus, it is necessary to investigate the absorption performance of the terahertz absorber at different incident angles. We simulated the incident angle dependencies of the device using CST Microwave Studio. The incident angles were varied from 0 to 75°. The simulation results from 0.1 THz to 10 THz are illustrated in Fig. 5, where the Fig. 5(a) is the absorption of the device with designed thicknesses, and Fig. 5(b) is the results using the real thicknesses. To complete the simulation from 10 THz to 20 THz, a more powerful computational server is essential, which is out of our ability so far. And the properties in this range are similar with that in Fig. 5. From the simulation, we can see that as the incident angle increases, the absorption for both the designed and real devices decreases rapidly at the low frequency range of 0.1–1 THz. For frequencies larger than 1.0 THz, both devices show insensitivity to the incident angle between 0 and 75°. The absorptions exceed 98% and 95% for designed and real devices, respectively. At the low frequency range < 1.0 THz, the decrease in the absorption is mainly due to the increasing reflectance as the incident angle increases according to the Fresnel formula. Moreover, the thickness of each composite layer also affects the AR effect. From the graded-index mechanism, the AR effect will decrease quickly when the thickness is far below the incident wavelength.
To evaluate the device scattering features, a terahertz QCL with a 4.2-THz center frequency and 200-GHz bandwidth was selected as an irradiation source. To detect the scattered terahertz wave, a Golay cell with a lock-in amplifier was utilized, and the output voltage of the lock-in amplifier was recorded as the scattering intensity. The output intensity of QCL is 139 mV. The Golay cell scan range was ± 20° from the reflective angle. The experimental results and setup are shown in Fig. 6. A coarse stainless plate was measured for comparison. From Fig. 6, for the stainless plate, the dependence of the scattering intensity and deviated angle follow a Gaussian dispersion (solid curve), and the maximum scattered value is 23.5 mV, which means approximately 16.90% of the incident THz laser was scattered at the reflective angle. The red dots in Fig. 6 are the scattering signal of our ultra-broadband perfect absorber. Compared with the stainless plate, our device exhibits very low scattering intensity and very weak angle dependency. The maximum scattered amplitude is only 0.47 mV. Note that the background noise level is approximately 0.1–0.2 mV; thus, only 0.19–0.27% can be the maximum scattering.
In summary, an ultra-broadband perfect absorber based on graded-index mechanism is designed and fabricated. To realize the refractive index profile gradient, the refractive index of the epoxy resin was widely tuned by hollow PS microspheres and TiO2 nanoparticles. The flat graded-index structure is particularly important for the ultra-broadband AR. With the flat six-layer structure and a heavily doped silicon substrate, 98% absorption can be achieved within the range of 0.1–20 THz. Moreover, the high absorption efficiency as well as the ultra-broad range is not limited by the incident angle, which maintains these features from 0 to 75° without an obvious decrease. The device has many advantages, such as ultra-broadband perfect absorption, low cost, and a convenient fabrication method allowing large-scale production. Our device has wide application potentials from thermal detectors, microbolometers, and imaging systems to terahertz stealth fields. Furthermore, according to Kirchhoff’s law of thermal radiation, the absorptivity of a material is equal to its emissivity at equilibrium. The multilayer AR device, therefore, can also be utilized to cover terahertz sources or terahertz blackbody devices.
Major National Development Project of Scientific Instrument and Equipment (2012YQ14000504); National Natural Science Foundation of China (NSFC) (1377111, and 61306118); Foreign Researcher Invitation Program of the National Institute of Information and Communications Technology, Japan.
The authors wish to thank Prof. H. Li of Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences for his experimental support of the scattering measurement and helpful discussions.
References and links
6. H.-T. Chen, J. Zhou, J. F. O’Hara, F. Chen, A. K. Azad, and A. J. Taylor, “Antireflection coating using metamaterials and identification of its mechanism,” Phys. Rev. Lett. 105(7), 073901 (2010). [CrossRef] [PubMed]
7. J. Hao, J. Wang, X. Liu, W. J. Padilla, L. Zhou, and M. Qiu, “High performance optical absorber based on a plasmonic metamaterial,” Appl. Phys. Lett. 96(25), 251104 (2010). [CrossRef]
8. W. Zhu and X. Zhao, “Metamaterial absorber with dendritic cells at infrared frequencies,” J. Opt. Soc. Am. B 26(12), 2382–2385 (2009). [CrossRef]
10. 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]
12. Q. Ye, Y. Liu, M. Li, and H. Yang, “Multi-band metamaterial absorber made of multi-gap SRRs structure,” Appl. Phys., A Mater. Sci. Process. 107(1), 155–160 (2012). [CrossRef]
13. X. Y. Peng, B. Wang, S. Lai, D. H. Zhang, and J. H. Teng, “Ultrathin multi-band planar metamaterial absorber based on standing wave resonances,” Opt. Express 20(25), 27756–27765 (2012). [CrossRef] [PubMed]
14. C. Shi, X. F. Zang, Y. Q. Wang, L. Chen, B. Cai, and Y. M. Zhu, “A polarization-independent broadband terahertz absorber,” Appl. Phys. Lett. 105(3), 031104 (2014). [CrossRef]
15. X. Zang, C. Shi, L. Chen, B. Cai, Y. Zhu, and S. Zhuang, “Ultra-broadband terahertz absorption by exciting the orthogonal diffraction in dumbbell-shaped gratings,” Sci. Rep. 5, 8901 (2015). [CrossRef] [PubMed]
16. Y. W. Chen, P. Y. Han, and X.-C. Zhang, “Tunable broadband antireflection structures for silicon at terahertz frequency,” Appl. Phys. Lett. 94(4), 041106 (2009). [CrossRef]
17. D.-S. Kim, D.-J. Kim, D.-H. Kim, S. Hwang, and J.-H. Jang, “Simple fabrication of an antireflective hemispherical surface structure using a self-assembly method for the terahertz frequency range,” Opt. Lett. 37(13), 2742–2744 (2012). [CrossRef] [PubMed]
18. X. C. Wang, Y. Z. Li, B. Cai, and Y. M. Zhu, “High refractive index composite for broadband antireflection in terahertz frequency range,” Appl. Phys. Lett. 106(23), 231107 (2015). [CrossRef]
19. B. Cai, O. Sugihara, H. I. Elim, T. Adschiri, and T. Kaino, “A novel preparation of high-refractive-index and highly transparent polymer nanohybrid composites,” Appl. Phys. Express 4(9), 092601 (2011). [CrossRef]