Terahertz-related technologies have received a lot of attention, due to their possible use in various fields such as high-speed communication, spectroscopic/imaging systems, and medical applications [1-5]. Among those, extensive research has been conducted to develop optical devices, such as filters and modulators, working at terahertz (THz) frequencies [6-8]. Actively controlling THz waves with improved functionality is necessary to realize the potential of optical devices. Much research has been carried out to investigate optical control of the spatial location, polarization, phase, propagation direction, and amplitude of reflected and transmitted THz waves, using a variety of materials and structures [9-18]. In particular, active THz modulation based on organic/inorganic hybrid structures has been suggested for realizing high modulation efficiency, caused by the change in optical properties due to the dynamics of photoexcited carriers at the interfaces of the hybrid structures [16, 17].

Recently, experimentally tuning the resonant response of artificially structured materials (known as metamaterials) has become a topic of interest, due to the possibilities of developing active THz devices for specific requirements [19-21]. In particular, using dielectric thin films with subwavelength thicknesses, a remarkable modification of the resonance was theoretically proposed and experimentally observed [22, 23]. The underlying mechanism of the phenomenon is the degree of a dielectric thin layer’s contribution when interacting with metamaterial structures. To take advantage of this technique in practice, the characteristics of the resonance shifts of metamaterials should be investigated for a variety of materials and structures, and an effective method for actively controlling the resonance should be suggested.

In this paper we experimentally observe the resonance shifts of metamaterials in organic/metamaterial/silicon hybrid structures, by varying the subwavelength thickness of organic thin film, and by exciting carriers on a silicon substrate under photoexcitation. The transmission properties of organic/metamaterial/silicon structures with different layer thickness of tris (8-hydroxyquinolinato) aluminum (Alq₃) are measured using THz time-domain spectroscopy under externally excited laser irradiation. The dramatic shift of the resonance frequency due to the change in thickness of the Alq₃ layer indicates that the contribution of organic thin layers when interacting with the metamaterial/silicon structures is highly sensitive to the subwavelength thickness. In addition, the blueshift of the resonant frequency under photoexcitation shows that the fundamental inductive-capacitive resonance can be tuned by changing the effective capacitance generated inside the gaps of combined split-ring resonators (SRRs). Our simulation results also argue that the degree of resonant-frequency shift due to photoexcitation depends on each resonance mode.

Figure 1(a) shows the schematic of an experimental THz wave transmission through organic/metamaterial/silicon hybrid structures under laser light irradiation. The side view of a typical element shows the metamaterial inclusion structure, sandwiched by organic and silicon (Si) substrates, as schematically depicted in Fig. 1(b). All samples are fabricated on a 500 μm thick inorganic layer of highly resistive Si (1.0 × 10⁵ Ω cm). The metamaterial structures are deposited on the Si substrate using a conventional photolithographic technique, forming metal patterns of Cr and Au (2 nm/150 nm approx.). Figure 1(c) shows the geometry and dimensions of a THz metamaterial’s unit cell. The unit element consists of two single SRRs, back to back. The two gaps formed on the sides have a fixed gap size a of 2 μm and a length b of the metallic bar of 7 μm. The width of all wires w is 4 μm. The side lengths m and n of the square unit element are 76 μm. The basic elements are arranged with periods P_x and P_y of 100 μm throughout the whole metamaterial area of 10 × 10 mm². The metamaterial layer is used as a frequency-selective surface, selectively transmitting incident THz waves at specific frequencies. The resonant response is determined by the fundamental LC resonance due to the values of an inductance L and a capacitance C, which are provided by the structural dimensions of the SRRs. Here a net electric response becomes prominent, since the two symmetric inductive loops make the magnetic response negligible.

The third layer shown in Fig. 1(b) is a thin organic layer, consisting of Alq₃ molecules (99.9%) from Organic Semiconductor Materials (OSM) [24, 25]. To fabricate an ultrathin dielectric substrate on the top of the metamaterial/silicon structures, the Alq₃ molecules are deposited using a vacuum thermal evaporation method (10⁻⁶ torr) at room temperature. Alq₃ layers with different thicknesses (t₁ = 0, 1, 5, and 10 μm) are fabricated by controlling the evaporation rate.

The transmittance of THz waves is measured using a typical THz time-domain spectroscopy system [26-28]. The optically induced THz radiation from a biased photo-conductive antenna transmits through the samples, and is detected by a THz photoconductive antenna. The polarization direction of incident THz waves is set to zero at the configuration shown in Fig. 1(a). We obtain the transmitted time-domain waveforms with (or without) photoexcitation, which is generated using a CW diode laser of wavelength 785 nm and intensity 200 mW. The diameters of the focused THz waves and optical pump beams are consistently ~3 mm. The THz spectra in the range of 0.1-1.2 THz are obtained by applying a fast Fourier transform method to the measured time-domain waveforms.

Figure 2 shows the measured transmission spectra for four samples with different thicknesses, obtained without Fig. 2(a) and with Fig. 2(b) photoexcitation. The polarization of the incident THz waves is parallel to the wires in the middle. Two distinct resonances are clearly observed in both cases. Whether an optical beam impinges on the surface of the samples or not, the resonances show frequency-dependent response as a function of thickness t₁, which once more verifies the previous results showing the change of electric response on ultrathin dielectric substrates [20, 23]. When the Alq₃ layer is absent from the metamaterial/silicon hybrid structure, the resonant peaks are located in the relatively highest-frequency region. With increasing thickness of the Alq₃ layer upon the metamaterial/silicon hybrid structure, the resonant peaks shift continuously toward lower frequencies.

First, to understand the origin of the two resonances, simulations of the structures are carried out at the two spectral peaks of 0.23 and 0.96 THz. We numerically investigate the near-electric-field distribution using COMSOL Multiphysics software. The simulations are performed on the metamaterial structures surrounded by perfectly matched layers in the direction perpendicular to the surface, and with periodic boundary conditions in the planes of the films. Figure 3 shows the simulated electric-field distributions obtained at the two resonant peaks on the top surfaces of the samples.

Each metamaterial unit cell is structurally symmetric about the wire in the middle. As a result, the center of the structures (marked by point A) acts as an electric-field node for both cases of the two resonances. On the other hand, the induced capacitance at the SRR gaps enables the flow of displacement current through the edges of the metallic wires, acting as electric-field antinodes. The resonance at 0.23 THz is therefore related to the fundamental inductive-capacitive response induced by circulating currents on the gross metallic area of each unit element. The electric field distribution at 0.96 THz shows three distinct nodal points, marked A, B, and C. The resonance can be regarded as the next-higher resonance mode generated by the change in capacitance at the gaps.

To reliably assure the change in spectral characteristics of the resonances of metamaterials, the values of spectral resonances are extracted and plotted as a function of the thicknesses of the organic thin layers for both cases, with and without photoexcitation, as shown in Fig. 4. The resonance shift to lower frequencies strongly supports the assertion that the effective substrate index of the organic thin layer stands out with increasing thickness of the organic film [22]. Note that all resonant peaks shift to higher frequencies under photoexcitation. The unit cell of the metamaterial composed of SRRs shows two metallic loops and two air gaps. The loops and gaps generate the loop inductance L and the gap capacitance C respectively. In general, the fundamental resonant frequency of the metamaterials is therefore given approximately by the inductive-capacitive resonance (LC resonance), [29-31]. The values of L and C can be varied with the structural dimensions of the total length of the metallic loops and the width of the air gap. This means that the resonant frequency due to the LC response can be controlled by properly designing the structure of the SRRs. In addition, change in the properties of surrounding materials leads to change in values of gap capacitance and loop inductance. The photo-excited carriers (generated at the substrate by the impinging optical beam) reduce the capacitive response associated with the SRR gaps. On the basis of the fundamental LC resonance, the reduction of the capacitive response leads to the blueshift of the resonance.

Next, we make a brief comparison between the results demonstrated above and the experimental data measured under x-axis polarization of incident THz waves, shown in Figs. 5(a) and 5(b). Figure 5(c) shows the extracted values of the spectral resonances observed under the condition of x-axis polarization of the THz waves. This clearly shows that, as for the results measured under y-axis polarization, the resonance value shifts to higher frequency with decreasing thickness of the organic thin film on the metamaterial structures. In contrast, the blueshift phenomenon of the resonances with photoexcitation is not relatively clear, compared to the results shown in Fig. 4. This implies that the change of the capacitive response inside the SRR gaps scarcely contributes to change in spectral characteristics of the local resonance due to the metamaterial structure. This is due to the fact that the resonance appearing at 0.43 THz originates from the electric-dipole resonance occurring in the straight metal wires lying along the x-axis. In this situation, the SRR gaps would have difficulty functioning as local capacitors.

In conclusion, we demonstrate that the local resonances in organic/metamaterial/Si hybrid structures can be tuned not only by varying the thickness of the organic thin film, but also by controlling the photoexcitation condition. The redshift of the resonances occurring when the thickness of the organic thin layer increases (on the subwavelength scale) is associated with the effect of a thin dielectric substrate. The blueshift of the resonances under photoexcitation is due to the reduction of capacitive response by the photo-excited carriers generated at the substrate. This is explained on the basis of the fundamental inductive-capacitive resonance. These results offer invaluable insights into exploiting metamaterial-based hybrid structures for developing optical systems requiring refractive-index engineering and active optical devices.