Research on the Terahertz Modulation Performance of VO₂ Thin Films with Surface Plasmon Polaritons Structure

Abstract

This paper focuses on the switching and modulation techniques of terahertz waves, develops ${\mathrm{VO}}_2$ thin-film materials with an SPP structure, and uses terahertz time-domain spectroscopy (THz-TDS) to study the semiconductor–metal phase transition characteristics of ${\mathrm{VO}}_2$ thin films, especially the photoinduced semiconductor–metal phase transition characteristics of silicon-based ${\mathrm{VO}}_2$ thin films. The optical modulation characteristics of silicon-based ${\mathrm{VO}}_2$ thin films to terahertz waves under different light excitation modes, such as continuous light irradiation at different wavelengths and femtosecond pulsed laser irradiation, were analyzed. Combining the optical modulation characteristics of silicon-based ${\mathrm{VO}}_2$ thin films with the filtering characteristics of SPP structures, composite structures of ${\mathrm{VO}}_2$ thin films with metal hole arrays, composite structures of ${\mathrm{VO}}_2$ thin films with metal block arrays, and silicon-based ${\mathrm{VO}}_2$ microstructure arrays were designed. The characteristics of this dual-function device were tested experimentally. The experiment proves that the ${\mathrm{VO}}_2$ film material with an SPP structure has a transmission rate dropping sharply from 32% to 1% under light excitation; the resistivity changes by more than six orders of magnitude, and the modulation effect is remarkable. By applying the SPP structure to the ${\mathrm{VO}}_2$ material, the material can simultaneously possess modulation and filtering functions, enhancing its optical performance in the terahertz band.

1. Introduction

Terahertz radiation (0.1–10 THz) occupies the spectral region between microwaves and infrared waves. Since the late 20th century, it has emerged as a significant frontier in scientific research [1,2,3]. Advances in femtosecond laser technology, together with the development of efficient terahertz sources and detectors [4], have collectively driven breakthroughs in key enabling technologies. ${\mathrm{VO}}_2$, a prototypical phase-change material, has attracted extensive research interest due to its abrupt conductivity change during the insulator-to-metal transition (IMT). Pioneering studies since the year 2000 have demonstrated that thermal excitation [5], optical excitation [6,7], and electrical excitation [8] can effectively modulate terahertz transmission by more than three orders of magnitude, providing ideas for the development of terahertz devices.

The phase transition of ${\mathrm{VO}}_2$ leads to a significant change in its electrical conductivity [9,10,11]. This characteristic can be utilized for the amplitude modulation of terahertz waves, and the periodic SPP (surface plasmon polariton) structure exhibits obvious frequency selectivity [12,13]. By combining these two points, two-dimensional thin-film materials with good modulation performance for terahertz waves can be prepared. These thin-film materials can be used in areas such as terahertz detection, terahertz imaging, terahertz optical switches, and modulators [14,15,16,17,18,19,20].

In 1959, Morin first discovered the phase transition in the ${\mathrm{VO}}_2$ material system: under the excitation of light, heat, etc., ${\mathrm{VO}}_2$ undergoes a reversible phase transition from the semiconductor phase to the metallic phase [21]. This property gives ${\mathrm{VO}}_2$ great potential for applications in smart windows, photoelectric switches, and other fields. Changes in the crystal structure that accompany ${\mathrm{VO}}_2$ during phase transitions will also significantly affect the characteristics of its electronic band structure. Figure 1 shows the band structure of ${\mathrm{VO}}_2$ in phase. In the metallic phase of ${\mathrm{VO}}_2$, the $3d$ orbitals of ${\mathrm{V}}^{4+}$ and the $2p$ orbitals of ${\mathrm{O}}^{2-}$ undergo orbital hybridization, resulting in a wider $\pi$ bonding orbital and a narrower ${\pi }^{*}$ bonding orbital. Although the overall bandwidth is up to 2.5 eV, the orbitals $d_{\left|\right|}$ are only partially filled, resulting in a partial overlap with the $\pi$ band, and the Fermi level is precisely in the center of this overlapping band, which is the fundamental reason for the metallic conductivity of the material. When ${\mathrm{VO}}_2$ is in the semiconductor phase, the tilt of the the [${\mathrm{VO}}_6$] octahedron improves the hybridization intensity of the $\pi$ orbitals, causing the ${\pi }^{*}$ band to shift upwards. At the same time, the V atom moves, causing the $d_{\left|\right|}$ orbitals to split, forming the occupied state $d_{\left|\right|}$ and the non-occupied state $d_{\left|\right|}^{*}$. At this point, the ${\pi }^{*}$ band is completely above the Fermi level and becomes an empty conduction band, while the $d_{\left|\right|}$ band exists as a full-valence band, presenting typical semiconductor characteristics [22].

Related research has found a process called surface plasmon polarization (SPP) in metal subwavelength periodic structures, and the above terahertz filtering and bunging phenomena are closely related to SPP. Surface plasmonic polarization waves are a type of electromagnetic wave mode that can confine electromagnetic waves to the surface of metals. In this mode, the electromagnetic waves have the potential to be confined within subwavelength intervals, thus breaking through the limitations of the diffraction limit and significantly improving the resolution of terahertz wave imaging. In addition to the strong electromagnetic confinement characteristics, the resonance frequency of the SPP can be controlled by processing periodic subwavelength structures, so subwavelength metal wire grids, periodic metal hole arrays, etc., have obvious frequency-selective characteristics, which have important applications in filter design. However, the loss of these devices is generally large, so they have not received much attention. Since Ebbesen et al. discovered the anomalous enhanced transmission of subwavelength periodic hole arrays in 1998 [23] and proposed that this anomalous enhanced transmission was the effect of surface plasmonic polarization waves, devices with such surface plasmonic structures have aroused widespread interest. Due to the limitations of the processing scale, these devices are relatively difficult to implement in the optical band, but in the terahertz band, subwavelength surface plasmonic structures are relatively easy to implement because the wavelength is much larger than that of the light wave, and many metal materials have properties close to ideal conductors in the terahertz band. The loss of the SPP in the terahertz band is very low when it travels over metal surfaces, so terahertz surface plasmon polarization waves have received particular attention [24,25].

In this paper, ${\mathrm{VO}}_2$ thin-film materials were fabricated on silicon-based metal SPP structures by means of magnetron sputtering. The physical and chemical composition of the materials was tested, and the optical properties of the ${\mathrm{VO}}_2$ materials on the SPP structure in the terahertz band were investigated using THz-TDS. The optical transmission, modulation, and frequency modulation characteristics in the terahertz band were studied.

2. Experiment

2.1. Preparation of ${\mathit{VO}}_{\mathit{2}}$ Thin Films

In this paper, a single ${\mathrm{VO}}_2$ thin film was first fabricated on a silicon substrate by magnetron sputtering [26,27], and the physicochemical composition of the ${\mathrm{VO}}_2$ thin-film material and its optical transmission behavior in the terahertz band were tested. In this paper, P-type silicon single-crystal polished wafers with a crystal orientation of <100>, $\rho >5000\mathrm{\Omega }$/cm, and a thickness of 425 $\mathsf{\mu }$m were used as substrates; the sputtering process parameters were $Ar$:$O_2$ = 48:0.8 (sccm) and the working pressure was 1 Pa. In order to obtain ${\mathrm{VO}}_2$ material with significant phase transition characteristics, the process temperature was selected from room temperature to 550 °C to select the optimal ${\mathrm{VO}}_2$ material.

2.2. ${\mathit{VO}}_{\mathit{2}}$ Material with an SPP Structure

Two SPP-structured arrays were fabricated by photolithography and thermal evaporation processes. Design 1 was a silicon-based metal hole array covered with a ${\mathrm{VO}}_2$ thin film, and Design 2 was a silicon-based metal block array covered with a ${\mathrm{VO}}_2$ thin film.

The first type of design was achieved through the following process: First, the silicon-based metal hole array was fabricated by photolithography and thermal evaporation, and the evaporated metal was Al with a thickness of about 500 nm. The ${\mathrm{VO}}_2$ thin film, 600 nm thick, was then covered on the sample surface by magnetron sputtering. Two samples of this type were prepared in the experiment with the following parameters:

Sample 1: Period 160 $\mathsf{\mu }$m, two-dimensional regular lattice structure, metal hole length 100 $\mathsf{\mu }$m, width 80 $\mathsf{\mu }$m;

Sample 2: Period 100 $\mathsf{\mu }$m, two-dimensional square lattice structure, metal hole length 50 $\mathsf{\mu }$m, width 25 $\mathsf{\mu }$m;

The microscopic image of the physical object is shown in Figure 2.

The second type of design was implemented in the same way as the first one. Three different parameters were also prepared in the experiment:

Sample 3: Period 160 $\mathsf{\mu }$m, two-dimensional regular lattice structure. Metal block length 100 $\mathsf{\mu }$m width 80 $\mathsf{\mu }$m;

Sample 4: Period 100 $\mathsf{\mu }$m, two-dimensional square lattice structure. Metal block length 50 $\mathsf{\mu }$m width 25 $\mathsf{\mu }$m;

A microscopic image of the object is shown in Figure 3.

2.3. Terahertz Optical Experiments

In this paper, terahertz time-domain spectroscopy technology was adopted to conduct relevant optical experiments [28,29]. The typical THz-TDS measures the time-domain transmitted signal of the sample and obtains spectral information through Fourier transform [30,31]. Therefore, in addition to obtaining the spectrum of the sample, THz-TDS can also obtain phase information that is not available with ordinary spectrometers, which is the most significant feature of THz-TDS. The experimental setup is shown in Figure 4. This system uses a titanium–sapphire femtosecond laser as the pumping source, with a central working wavelength of 800 nm, a repetition frequency of 100 MHz, a pulse width of 50 fs, and an average power of 520 mW. The generation and detection of terahertz waves are carried out through micro-optical rectification and electro-optic sampling, and the nonlinear crystal used is zinc telluride (ZnTe). The effective spectral range of the system is 0.3 to 2.5 THz. The spot size on the ${\mathrm{VO}}_2$ thin film surface is controlled by the telescope system composed of lenses L1 and L2. Through multiple experiments and optimizations, it was found that the best effect is achieved when the spot diameter is 0.4 mm.

3. Experimental Results and Discussion

3.1. Composition and Surface Morphology Analysis of ${\mathit{VO}}_{\mathit{2}}$ Nanoparticles

The crystal structure of the samples obtained at different heat treatment temperatures was characterized by X-ray diffraction (XRD) to analyze the variation of the crystal structure with heat treatment temperature. The test results are shown in Figure 5. Distinct characteristic diffraction peaks at 2$\theta$ = 27.80°, 37.09°, 42.27°, 55.45°, 57.42°, 65.00°, 70.25°, and 72.02° were observed in the XRD pattern of the unheat-treated sample S1. According to JCPDS PDF card 09-0142, these diffraction peaks all originated from the monoclinic structure ${\mathrm{VO}}_2$, corresponding to the (011), (200), (210), (220), (022), (013), (231), and (−411) crystal directions, respectively. The main characteristic diffraction peaks of the heat-treated samples were similar to those of the unheat-treated sample S1, indicating that the main crystal phase of the six samples remained M-phase ${\mathrm{VO}}_2$ throughout the heat treatment process. Figure 5 also shows a small number of diffraction peaks corresponding to ${\mathrm{V}}_2{\mathrm{O}}_5$ in the 350 °C heat-treated sample S2, which may be the result of the oxidation of ${\mathrm{VO}}_2$; as the heat treatment temperature gradually increased, the impurity peaks associated with ${\mathrm{V}}_3{\mathrm{O}}_7$ disappeared, and diffraction peaks associated with ${\mathrm{V}}_2{\mathrm{O}}_5$ appeared, with increasing intensity. Despite the formation of impurities during heat treatment, the main component of most samples remained ${\mathrm{VO}}_2$.

In order to study the influence law of the heat treatment temperature on the composition and valence state of ${\mathrm{VO}}_2$, X-ray photoelectron spectroscopy (XPS) tests were conducted on six samples. The surface XPS spectra of the V$2p$ state are shown in Figure 6, and the peaks were normalized based on the 284.8 eV binding energy of C1s. The characteristic peaks of V$2p_{1/2}$ and V$2p_{3/2}$ can be clearly observed in Figure 6 due to orbital splitting [32]. The two characteristic peaks were fitted within the binding energy ranges of ${\mathrm{V}}^{5+}$(516.9 eV–517.7 eV), ${\mathrm{V}}^{4+}$(515.7 eV–516.2 eV), and ${\mathrm{V}}^{3+}$(515.2 eV–515.9 eV) to obtain their component ratios [33]. The reduction of ${\mathrm{V}}^{4+}$ ions from ${\mathrm{V}}^{3+}$ to V is accompanied by electron transfer. The main chemical reactions that occur in a vacuum are as follows [34]:

$$\begin{array}{c}{O}_o^x\leftrightarrow v_O^{2+}+2e^{-}+{\displaystyle \frac{1}{2}}{O}_2\left(g\right)\end{array},$$

(1)

$$\begin{array}{c}{V}^{4+}+e^{-}\leftrightarrow V^{3+}\end{array}$$

(2)

Among them, $O_o^x$ represents the oxygen occupying the anion site, $v_O^{2+}$ represents the oxygen vacancy, and ${\mathrm{V}}^{4+}$ and ${\mathrm{V}}^{3+}$ respectively represent the original Vanadium ions in the state and reduced state. These chemical equations suggest a significant increase in the concentration of ${\mathrm{V}}^{3+}$ after heat treatment, which is associated with the formation of oxygen vacancies. Two sharp peaks were observed at 516.5 eV and 515.3 eV, representing the excitation of V ions in the +4 and +3 oxidation states at the V$2p_{3/2}$ energy level, respectively [33]. Based on the area percentages of different valence states of V ions after V$2p_{3/2}$ peak fitting, the relative concentrations of ${\mathrm{V}}^{3+}$, ${\mathrm{V}}^{4+}$, and ${\mathrm{V}}^{5+}$ in six samples were calculated, and the results are shown in

Table 1. The different ratios of the valence states of the V element in samples S1, S2, S3, S4, S5, and S6.

Samples	${\mathbf{V}}^{\mathbf{3}\mathbf{+}}$ (%)	${\mathbf{V}}^{\mathbf{4}\mathbf{+}}$ (%)	${\mathbf{V}}^{\mathbf{5}\mathbf{+}}$ (%)
S1	45.60	47.34	7.06
S2	60.85	31.78	7.37
S3	64.69	31.72	3.59
S4	63.00	36.85	0.15
S5	61.80	36.05	2.15
S6	60.31	35.53	4.16

. The concentration of ${\mathrm{V}}^{3+}$ ions demonstrates a progressive increase with elevating heat treatment temperature, attaining a peak value at 400 °C, beyond which a gradual decrease is observed. The oxygen vacancy concentrations in the five heat-treated samples (S2-S6) were represented by the relative increase in ${\mathrm{V}}^{3+}$ compared to the untreated sample S1, with values of 15.25%, 19.09%, 17.4%, 16.2%, and 14.71%, respectively. These results reveal that the vacuum heat treatment temperature provides a controllable means to engineer oxygen vacancy defects with tunable concentrations in ${\mathrm{VO}}_2$.

For the polycrystalline ${\mathrm{VO}}_2$ thin films, their electrical properties are not only related to the grain boundaries of the polycrystalline films, but also to the oxygen vacancies in the films. The existence of grain boundaries is equivalent to introducing impurity energy levels in the band gap, reducing the activation energy; oxygen vacancies introduce acceptor energy levels in the band gap, also resulting in a decrease in the activation energy [35]. Both types of defects will reduce the phase transition performance of the ${\mathrm{VO}}_2$ thin films. Moreover, oxygen vacancies will introduce excess electrons in the films, reducing the optical transmittance of the ${\mathrm{VO}}_2$ thin films. Furthermore, due to the relatively different oxidation states of vanadium in the ${\mathrm{VO}}_2$ thin films, the charging states of oxygen vacancies also differ [36]. Therefore, appropriate preparation process conditions are conducive to obtaining controllable grain distribution and oxygen vacancy states, thereby obtaining the optimal ${\mathrm{VO}}_2$ thin-film material for phase transition.

3.2. Test for Terahertz Optical Properties of ${\mathit{VO}}_{\mathit{2}}$ Thin Films

The experimental setup utilized a commercial 976 nm semiconductor laser from n-Light with a maximum output power of 40 W. We intentionally operated at reduced power levels due to both material constraints and practical considerations. Excessive laser power induces thermal expansion in the ${\mathrm{VO}}_2$ thin films, resulting in surface cracking and potential delamination. Moreover, higher modulation powers would compromise the system’s applicability for real-world implementations. Therefore, all measurements were conducted with the laser power maintained below 2 W.

The system was the same as before, but the light source was replaced with a semiconductor laser with fiber output, and the size of the laser spot on the VOx surface was the same as before, namely 0.4 mm.

Figure 7a,b show the terahertz pulses and their corresponding spectra measured at different modulated optical powers. Apparently, at a low power (<400 mW) modulation, the modulation of terahertz pulses is similar to that of 520 nm laser modulation, with no significant changes in the time-domain waveforms and spectral shapes of the pulses. However, at high power, the terahertz pulses are severely distorted, with a significant distortion of the spectral shape. This indicates that there is a significant difference in modulation corresponding to different frequency components.

Figure 8a shows the dependence of the modulation depth on the modulated optical power and carrier frequency. It can be seen from the figure that under the same optical power, the modulation of terahertz waves with different frequency components is significantly different. When the excitation light power is high, the modulation of the low-frequency components is close to 100%, almost saturated, while the high-frequency components still have room for further modulation. Because of this frequency-dependent characteristic, after the terahertz pulse passes through the ${\mathrm{VO}}_2$ thin film, the proportion of different frequency components is out of balance, resulting in a distortion of the spectral shape, which in turn causes a distortion of the time-domain waveform of the pulse.

Figure 8b shows the modulation depth–modulation optical power curve at 1 THz. When the modulated light intensity is high, the modulation tends to saturation. This is mainly because when the modulated light intensity is high, almost all phase-transitioning lattices in the silicon-based ${\mathrm{VO}}_2$ thin film are phase-transitioned, and no further increase in light intensity can produce more independent electrons, and the electrical conductivity cannot be further increased. Therefore, higher power modulation is of little significance.

3.3. Modulating Terahertz Waves with a Femtosecond Laser Pulse with a Central Wavelength of 1040 nm

With all of the above utilized CW lasers acting on ${\mathrm{VO}}_2$ thin films, and at the same average power, the photon number density of femtosecond lasers was much higher than that of CW lasers. Therefore, when femtosecond pulses irradiate ${\mathrm{VO}}_2$ thin films, in addition to phase transitions caused by single-photon absorption, there may also be multi-photon absorption processes, suggesting that the modulation efficiency of femtosecond lasers as modulated light should be higher than that of CW lasers. To verify our hypothesis, a femtosecond laser with a central wavelength of 1040 nm was used to irradiate silicon-based ${\mathrm{VO}}_2$ thin films, and its modulation effect on terahertz waves was investigated.

The experimental setup remained identical to that described previously, except that the light source was replaced with a lab-built photonic crystal fiber femtosecond laser amplifier. This system features a center wavelength of 1040 nm, a repetition rate of 50 MHz, a pulse width of 50 fs, and an output power capability of up to 20 W. To prevent potential thin film damage under high-power excitation, the laser power was still maintained below 2 W.

Figure 9a shows the terahertz time-domain signals transmitted at the excitation power measured at different powers, and Figure 9b shows the corresponding terahertz spectrum. At low power, the modulation effect of the femtosecond laser is similar to that of the CW mode. The transmitted pulses maintain their original time-domain and frequency-domain shapes, and at high power, intense pulse distortion also occurs.

Figure 10a shows the modulation depth at different femtosecond laser powers and frequencies. Compared with the modulation characteristics of the CW laser at 976 nm (Figure 8a), the modulation of the femtosecond laser is larger with the same modulation power and the same terahertz frequency component. Furthermore, at the same modulation power, the modulation depth at high frequencies is smaller than that at low frequencies, which is similar to the effect of CW lasers.

Figure 10b shows the relationship between the modulation depth and modulated optical power at 1 THz. Apparently, with the increase in laser power, the modulation shows signs of saturation very quickly, while at 976 nm CW laser modulation, the saturation is much slower (Figure 8b), which also proves that the modulation efficiency of femtosecond lasers is indeed higher than that of CW.

3.4. Metallic Property Analysis of Photoinduced Phase Change Silicon-Based ${\mathit{VO}}_{\mathit{2}}$ Thin Films

In the previous section, we conducted optical modulation experiments on terahertz pulses using silicon-based ${\mathrm{VO}}_2$ thin films and achieved relatively ideal modulation results. We also found that the phase transition of silicon-based ${\mathrm{VO}}_2$ thin films was not sensitive to the wavelength of the excitation light. Based on the results of the 1040 nm optical modulation experiment, we will further analyze the metallic properties of silicon-based ${\mathrm{VO}}_2$ thin films after phase transition.

When a terahertz wave passes through a phase-change thin film with a much smaller thickness than the wavelength, there exists a direct relationship between the complex transmittance ratio before and after the phase transition and the complex conductivity after the transition. Using the thin-film approximation, the relative transmittance can be approximately expressed as [37,38,39,40,41]:

$${\displaystyle \frac{T_{pump}\left(\omega \right)}{T_{ref}\left(\omega \right)}}=\left|{\displaystyle \frac{T_{pump}\left(\omega \right)}{T_{ref}\left(\omega \right)}}\right|e^{i\phi \left(\omega \right)}={\displaystyle \frac{1+n_s}{1+n_s+Z_0\tilde{\sigma }\left(\omega \right)d}}$$

(3)

$T_{ref}$($\omega$) and $T_{pump}$($\omega$) are the terahertz transmittance before and after the phase transition, respectively, which can be obtained from the Fourier transform spectra of the transmitted signals. Here, $n_s$ is the refractive index of the substrate (silicon), $Z_0$ is the vacuum wave impedance, and d is the thin film thickness. The complex conductivity of the thin film can then be expressed in terms of the amplitude and phase functions of the transmittance as follows:

$$\tilde{\sigma }\left(\omega \right)={\displaystyle \frac{1+n_s}{Z_{0}d}}({\displaystyle \frac{cos\phi \left(\omega \right)}{\left|T_{pump}\left(\omega \right)/T_{ref}\left(\omega \right)\right|}}-1-i{\displaystyle \frac{sin\phi \left(\omega \right)}{\left|T_{pump}\left(\omega \right)/T_{ref}\left(\omega \right)\right|}})$$

(4)

The advantage of THz-TDS over other spectral techniques is that it can obtain phase information in addition to the amplitude information of transmittance, thus avoiding the use of complex Kramers–Kronig relations [42] to calculate the complex conductivity and complex refractive index of the thin film.

The calculated complex conductivities at different excitation light powers are shown in Figure 11. It can be seen that with the increase in the excitation light power, the real part of the conductivity gradually increases and comes closer to the range of common metal conductivities (>${10}^6$ S/m).

Another strange phenomenon was found in the experiment: when the optical power increased to 2.098 W, two peaks appeared in the relatively flat conductivity curve near 0.8 THz and 1.5 THz. This suggests that the silicon-based ${\mathrm{VO}}_2$ thin film has additional loss at these two frequencies. This resonance-like absorption peak, which we analyzed, is due to the heterogeneity of the microstructure of the ${\mathrm{VO}}_2$ thin film. During the magnetron sputtering process, ${\mathrm{VO}}_2$ gradually adheres to the substrate and forms different crystalline states. If the crystallization is not uniform, it is possible to form “particles” with different photo-induced phase transition properties. Under optical illumination, the post-phase-transition metallic states exhibit heterogeneity, wherein particles with stronger metallicity display resonance characteristics analogous to dipole oscillations, thereby exciting surface plasmon oscillations.

3.5. Terahertz Optical Test of ${\mathit{VO}}_{\mathit{2}}$ Thin Films with an SPP Structure

In the experiment, five samples were irradiated with a 976 nm laser at 500 mW, and their terahertz transmittance was measured before and after illumination. The polarization direction of the terahertz wave was aligned along the x-axis.

Sample 1: As can be seen from the transmittance of sample 1 before and after light irradiation (Figure 12), the introduction of the ${\mathrm{VO}}_2$ thin film did not change the resonance frequencies of the hole arrays and the metal block arrays, but the change in the transmittance amplitude is very obvious. From the curve of the modulation regime $|\mathrm{\Delta }T/T_0|$ ($T_0$ is the transmittance before illumination), it is known that the transmittance at the resonance is more seriously affected than that at other frequencies.

The resonant properties of surface plasmon polaritons (SPPs) are highly sensitive to the dielectric constant ($\epsilon$) of the medium, particularly in metallic materials. When a ${\mathrm{VO}}_2$ thin film undergoes a semiconductor-to-metal phase transition under optical excitation, not only does its electrical conductivity exhibit a significant jump, but its dielectric constant also undergoes substantial changes. This dramatically disrupts the SPP resonance. Consequently, in metallic hole arrays, the originally enhanced terahertz wave transmission via SPP resonance is severely compromised, leading to a sharp decrease in the transmission efficiency. This mechanism enables ${\mathrm{VO}}_2$ thin films to effectively modulate specific terahertz frequency components when integrated with metallic SPP structures.

The process by which the resonance of the SP is disrupted can be directly observed by simulating the distribution of the electromagnetic field before and after illumination. Figure 12 shows the distribution of the electric field before and after the illumination of simulated sample 1. The ${\mathrm{VO}}_2$ thin film in the simulation was characterized by electrical conductivity, and the conductivity before the phase transition was selected based on Liang’s phase transition measurement of ${\mathrm{VO}}_2$ in 2021 [32], which was ${10}^2$ S/m. The conductivity after phase transformation is ${10}^4$ S/m. From the distribution of the electric field at the resonant frequency, it can be seen that the binding of the inner wall of the metal hole to the electromagnetic field is significantly reduced after the phase transition, and the intensity is also much lower than before the phase transition, indicating that the ${\mathrm{VO}}_2$ thin film does indeed disrupt the resonant conditions of the SP.

Sample 2: The principle of sample 2 is similar to that of sample 1, changing the period and size of the sample, and its resonant frequency is 0.83 THz. The results of the experiment and the simulated electric field distribution are shown in Figure 13.

Samples 3 and 4: The principle of samples 3 and 4 are different from those of samples 1 and 2. They achieve terahertz modulation by destroying the DLSP resonance [43]. Figure 14 and Figure 15 correspond to the experimental results and simulated electric field distribution of sample 3 and sample 4, respectively. The polarization direction of the incident terahertz is along the y direction.

From the graph, we can see that the transmittance of most frequencies decays to varying degrees after illumination; however, at the resonant frequency, the terahertz transmittance increases significantly. Because DLSPs exhibit resonant reflection properties—that is, when there is no light, the DLSPs excited by the metal block resist terahertz transmission at the resonant frequency—when exposed to light, the DLSP’s blocking effect on terahertz waves is significantly weakened because the resonance conditions of the DLSP are disrupted by the phase-changed ${\mathrm{VO}}_2$ thin film. As a result, despite the same attenuation caused by the phase transition of ${\mathrm{VO}}_2$ (e.g., for other frequency components), the overall transmission rate of the terahertz component at the resonant frequency still increased. Thus, we obtained a distinct property from samples 1 and 2, that is, the attenuation of the ${\mathrm{VO}}_2$ thin film instead brought an increase in terahertz transmittance. The simulated electric field distribution also fully reflects the disruption of DLSP resonance.

Another notable point is that the results show that the phase transition of ${\mathrm{VO}}_2$ has a significantly greater effect on DLSP than on SP, because in samples 1 and 2, the response peak of SP after illumination is still very obvious, but in sample 3, the resonance peak of the DLSP after illumination is almost completely flattened.

4. Conclusions

${\mathrm{VO}}_2$ thin films with thermally induced phase-transformation properties were prepared by magnetron reactive sputtering combined with rapid heat treatment. Through XPS and XRD analysis methods, it was found that ${\mathrm{VO}}_2$ with a monocline structure was the main component of the thin films. Fourier transform infrared spectroscopy tests showed that when the ${\mathrm{VO}}_2$ thin film undergoes phase transition, the terahertz wave transmittance undergoes a sudden change near the phase transition point. Under the excitation of a 976 nm pump light, the transmittance of the ${\mathrm{VO}}_2$ thin film drops sharply from 32% to 1%, and the switching characteristics are obvious. By combining the metal SPP structure with the ${\mathrm{VO}}_2$ thin-film material, the resonance frequency of the system was controlled by altering the SPP structure. In this paper, we draw on a phenomenological molecular model for analyzing composite SPP structures in recent years to conduct experimental and simulation analyses of the plasmon coupling process in the pore-block composite structure. The physical essence of the coupling concept in the model was explained by numerical simulation, providing a theoretical reference for the design of SPP filtering structures in the terahertz band.